How Volcanoes Work: a 25 Year Perspective 1888 2013

How volcanoes work: A 25 year perspective 1888 2013

CELEBRATING ADVANCES IN GEOSCIENCE Katharine V. Cashman† and R. Stephen J. Sparks School of Earth Sciences, University of Bristol, Bristol BS81RJ, UK Invited Review

ABSTRACT eruption triggers. Finally, we look at eruptions Hawaiian volcanism is associated with the themselves, from the ascent of magma through eponymous Hawaiian eruptive style, which is Over the past 25 years, our understanding the crust to the physical controls on eruption dominated by fl uid lava fl ows. Lava fl ows often of the physical processes that drive volcanic styles and generation of eruptive products. We emerge directly from dike-fed fi ssure systems; eruptions has increased enormously thanks recognize that it is impossible to do justice to all for this reason, shield volcanoes tend to be to major advances in computational and of these topics—or all of the scientists who have elongated along the fi ssure direction. Hawai- analytical facilities, instrumentation, and col- contributed to the contemporary understanding ian shield volcanoes are built of stacks of these lection of comprehensive observational, geo- of volcanism—in a single article. In this regard, fl ows; their low slopes refl ect both the fl uidity physical, geochemical, and petrological data we note that a thorough review of the fi eld was of the initial lava and the tendency for lava fl ows sets associated with recent volcanic activity. completed in 2000 with the publication of the En- to thicken (because of cooling, crystallization, Much of this work has been motivated by the cyclopedia of Volcanoes (Sigurdsson et al., 2000). and associated increases in viscosity) with trans- recognition that human exposure to volcanic port distance from rift zone vents (e.g., Katz and hazard is increasing with both expanding VOLCANIC ERUPTIONS— Cashman, 2003). An unprecedented look at populations and increasing reliance on infra- AN OVERVIEW the structure of Hawaiian volcanoes has been structure (as illustrated by the disruption provided by a 15-yr-long drilling project that to air traffi c caused by the 2010 eruption of Volcanoes vary greatly in morphology, evolu- recovered core from Hawaii’s Mauna Loa vol- Eyjaf jallajökull volcano in Iceland). Reducing tion, eruptive styles, and behaviors as a conse- cano to a depth of ~3500 m, which represents vulnerability to volcanic eruptions requires a quence of the wide variety of tectonic settings, an ~700 k.y. history of the Hawaiian plume thorough understanding of the processes that melt production rates, magma compositions, and (Stolper et al., 2009). Not surprisingly, given govern eruptive activity. Here, we provide eruption conditions that they represent. Here, the proximity of the drilling site to the current an overview of our current understanding we introduce common volcanic landforms, to- shoreline, subaerial lavas represent only a small of how volcanoes work. We focus particu- gether with the eruption styles responsible for fraction of the core samples, with most of the larly on the physical processes that modulate their formation. Because magma composition volcanic sequence represented by subaqueous magma accumulation in the upper crust, is an important control on eruptive style, we hyaloclastites and pillow basalts. transport magma to the surface, and control separate discussions of mafi c and intermediate/ From a hazards perspective, an important eruptive activity. silicic volcanism. discovery about Hawaiian volcanism has been the recognition that Kilauea volcano has expe- INTRODUCTION Mafi c Volcanoes rienced periods of highly explosive activity in addition to the effusive eruptions of the past Volcanic eruptions are a spectacular manifesta- Mafi c volcanoes vary greatly in scale and few centuries (Fiske et al., 2009). Episodes tion of a dynamic Earth. They not only link deep construction style. The iconic basaltic landform of explosive activity are particularly frequent Earth (the geosphere) to the hydrosphere, atmo- is a shield volcano, such as those that comprise during times immediately following summit sphere, and biosphere but also affect human popu- the Hawaiian Islands, United States (Fig. 1A). caldera formation (Swanson, 2008; Swanson lations: ~600 million people live close enough to Shield volcanoes are constructed primarily by et al., 2012). Summit calderas in mafi c shield an active volcano to be affected by eruptions, and successive lava fl ows and are commonly char- volcanoes form by rapid drainage of magma civilization itself could be threatened by the larg- acterized by relatively low slopes. Other mafi c from summit storage regions to fl ank vents (e.g., est explosive eruptions that have occurred in Earth volcano morphologies include the “inverted Gudmundsson, 1987). In Hawaii, this drainage history. The core questions of volcanology focus soup bowl” shape of Galapagos volcanoes; allows access of groundwater to the magmatic on how volcanoes work, that is, how magma steep-sided cones, like Pico volcano in the system, which may fuel the high explosivity ob- forms and moves to the surface, and how the spe- Azores and Kluchevskoi in Kamchatka; fi s- served in postcaldera periods. cifi c properties of the magma, and the lithosphere sure volcanoes in tectonic rifts such as Iceland, Stromboli volcano, Italy (Fig. 1B), is the through which it moves, control eruptive activity. where they may be associated with a central sub- type location for the Strombolian eruption style, Here, we review progress that has been made on sidence caldera; tuja volcanoes erupted under ice which is characterized by frequent (often sev- this core topic over the past quarter century. To or in shallow-marine environments; mid-ocean eral per hour) small explosions that have been provide a context, we start by reviewing volcanic ridges with morphologies that refl ect spreading attributed to the rise and bursting of large indi- landforms and associated styles of eruptive activ- rate; and fi elds of monogenetic volcanoes, each vidual gas bubbles (e.g., Vergniolle and Jaupart, ity. We then describe our current understanding of related to a single eruptive episode. These land- 1986). Stromboli thus represents an “open-sys- magma storage regions (magma chambers) and forms refl ect a range in eruptive styles, the most tem” volcano, that is, a volcano where gases can common of which are reviewed next (see also move freely through the system. In fact, Strom- †E-mail: [email protected] Francis et al., 1990). boli typically produces ~105 times more gas

GSA Bulletin; May/June 2013; v. 125; no. 5/6; p. 664–690; doi: 10.1130/B30720.1; 18 ﬁ gures.

A B

C D

Figure 1. Photographs of mafi c volcanic landforms. (A) Mauna Loa shield volcano, United States; (B) Stromboli stratovolcano, Italy; (C) Paricutin cinder cone, Mexico; (D) Colli Albani mafi c caldera complex, Italy, viewed from Rome. than can be accounted for by the magma ejected face; whether this characteristic requires inter- generation of large mafi c ignimbrite deposits is beyond the vent (e.g., Harris and Ripepe, 2007). action with external water sources remains a curious from several perspectives, including the However, Stromboli can also produce lava fl ows matter of debate (e.g., White and Ross, 2011). mechanisms by which large volumes of mafi c and “paroxysmal” eruptions, as demonstrated in Improving our understanding of small mafi c (and very low viscosity) magma accumulate in 2002–2003 and 2007 (e.g., Ripepe et al., 2005; eruptions is important because cities such as the upper crust (rather than rise to the surface Calvari et al., 2008; Scandone et al., 2009). This Auckland, New Zealand, Bend, Oregon, and in small batches) and maintain sustained ex- variability in eruption style derives from the Mexico City, Mexico, are constructed within plosive activity (rather than losing volatiles and complex structure of the magma storage and active cinder cone fi elds. changing to effusive eruption styles). transport system, and the resulting alternation More important for hazards, and more puz- Another exciting advance in mafi c volcanism between near-surface and deep controls on erup- zling from the perspective of physical vol- over the past few decades has come from the tive activity. canol ogy, is the recent documentation of highly oceans. Studies of submarine volcanism in- Cinder cone fi elds characterize regions of ac- explosive eruptions from mafi c volcanic cen- creased with the advent of the RIDGE program tive extension and transtension (Fig. 1C). Here, ters. Widespread tephra deposits from mafi c of the 1980s and 1990s, which greatly enhanced ascent and eruption of small mafi c magma volcanoes were fi rst recognized from eruptions our understanding of processes occurring in mid- batches produce a spectrum of eruptive styles of Masaya, Nicaragua (Williams, 1983; Bice, ocean-ridge environments. Mid-ocean ridges are from fi ssure-fed Hawaiian lava fl ows to Strom- 1985). Interest in mafi c Plinian eruptions re- sites of frequent volcanic activity that is typically bolian bubble bursts to explosive gas-charged vived with documentation of a mafi c Plinian manifested as fi ssure-generated lava fl ow erup- violent Strombolian eruptions to passive lava eruption from Etna volcano in 122 B.C. (Coltelli tions of varying intensities (e.g., Rubin et al., effu sion. Hawaiian-style eruptions are domi- et al., 1998) and has led to numerous detailed 2012). These eruptions can be monitored where nated by lava fl ows, Strombolian-style eruptions fi elds studies of mafi c explosive volcanism (for ocean-based hydrophone networks are suffi - produce small scoria cones and/or lava fl ows, example, Cas and Giordano, 2006; Pérez and ciently dense to record T-phase seismicity asso- and violent Strombolian eruptions produce sub- Freundt, 2006; Costantini et al., 2010). Most ciated with magma migration to the surface (e.g., stantial tephra sheets. Other features that are im- surprising, however, has been the recognition Slack et al., 1999). Axial volcano on the Juan de portant in the spectrum of small mafi c volcanoes of very large (tens of cubic kilometers) mafi c Fuca Ridge (NE Pacifi c) also hosts a sub- are maars and diatremes, which have craters that ignimbrites from Colli Albani volcano, Italy marine monitoring network of seismic, pressure , have excavated well below the pre-eruptive sur- (Fig. 1D; Funiciello and Giordano, 2010). The and deformation sensors that has now recorded

Geological Society of America Bulletin, May/June 2013 665 Cashman and Sparks multiple eruptions (e.g., Fox et al., 2001; Nooner tral cones are often surrounded by gently dip- seconds) but intense explosions characterize and Chadwick, 2009; Caress et al., 2012; Chad- ping fl anks composed of lavas and pyroclastic Vulcanian eruptions, which are most common in wick et al., 2012; Dziak et al., 2012; Mitchell, and volcaniclastic (particularly volcanic mud- volcanoes of intermediate (andesitic to dacitic) 2012). Additionally, recent remotely operated fl ow) material. Although not always interme- compositions. The type locality—Vulcano— vehicle (ROV) cruises to the western Pacifi c diate in composition (Etna, Italy, Fuji, Japan, is rhyolitic, but also erupts latite and trachyte have identifi ed several mafi c submarine arc vol- and Villarica, Chile, are basaltic examples), magmas. The continuum from sustained Plin- canoes that are either commonly or persistently this geomorphic form typifi es volcanoes con- ian through pulsatory subplinian to Vulcanian active (e.g., Embley et al., 2006). One of these, structed from viscous (often intermediate/ activity derives from variations in magma kinet- NW Rota-1, lies ~100 km north of Guam in the silicic) magmas that erupt explosively as well ics and dynamics during ascent (e.g., Cashman, Marianas, has been erupting since at least 2004, as effusively. 2004; Mason et al., 2006). and has produced eruptions that range from ef- Explosive eruptions of stratovolcanoes are Stratovolcanoes are prone to failure either fusive to mildly explosive (e.g., Chadwick et al., classifi ed as Plinian if they are large (with erup- by sector collapse, as illustrated by the 1980 2008; Fig. 2). Studies of these systems provide tion column heights in excess of 20–25 km and eruption of Mount St. Helens (Lipman and important insight into processes that form both dense rock volumes >1 km3), sustained, and pro- Mullineaux, 1981), or by caldera formation, as oceanic crust and the island-arc component of duce widespread tephra deposits (Newhall and occurred at Mount Pinatubo in 1991 (Newhall continents. Self, 1982). This term derives from the 79 A.D. and Punongbayan, 1996). Sector collapse is (Pompeii) eruption of Vesuvius, Italy, and can commonly accompanied by explosive activ- Intermediate/Silicic Volcanoes be applied to sustained eruptions of Mount St. ity; particularly lethal are resulting laterally di- Helens, United States, in 1980 and Pinatubo, rected blasts, which occur when the edifi ce (or Stratovolcanoes are perhaps the best-known Philippines, in 1991. Eruptions with smaller dome) fails because of intruding magma (e.g., (and most iconic) volcano type (Fig. 3A). They volumes (0.1–1 km3), lower eruption columns Druitt, 1992; Hoblitt, 2000; Voight et al., 2002). are steep-sided cones constructed from stacked (<20–25 km), and more local tephra deposits are Caldera collapse follows withdrawal of large lavas and pyroclastic deposits (Fig. 3B). Cen- termed subplinian. Short-lived (typically tens of volumes of magma. Famous caldera-forming

A B

14°40′

14°38′

14°36′ C

14°34′

14°32′N

144°42′E 144°44′ 144°46′ 144°48′ 144°50′

Figure 2. NW Rota-1 submarine volcano. (A) Bathymetry and location map; (B) explosion showing quench fragmentation within the plume; (C) red lava explosion (modiﬁ ed from Chadwick et al., 2008; Deardorff et al., 2011).

666 Geological Society of America Bulletin, May/June 2013 How volcanoes work

A B

C D

Figure 3. (A) Villarica stratovolcano, Chile. (B) Crater walls of Santorini volcano, Greece, showing interlayered lava fl ows and pyroclastic deposits. (C) Temple of Serapis, Pozzuoli, Italy. This site lies within the Campi Flegrei caldera and has experienced repeated uplift and subsidence, with the most recent uplift in the mid-1980s. (D) Lava spine from the 2004–2008 eruption of Mount St. Helens, Washington (U.S. Geological Survey photograph). eruptions include the ca. 7700 yr B.P. eruption commonly termed “supervolcanoes,” as they with these events derives in part from the recog- of Crater Lake, Oregon (Bacon, 1983; Bacon represent an extreme end member of volcanic nition that large caldera-forming eruptions may et al., 2002), the ca. 3600 yr B.P. eruption of San- activity. Caldera formation in these systems is have played a role in the development of human torini, Greece (Druitt et al., 1989), and the more attributed to subsidence related to rapid with- history, such as the 39,000 yr B.P. eruption of recent eruptions of the Indonesian volcanoes drawal of very large volumes of magma (tens Campi Flegrei, Italy (e.g., Fedele et al., 2008) Tambora (A.D. 1815; Oppenheimer, 2003) and to a few thousand km3) in single events; sub- and the 75,000 yr B.P. eruption of Toba, Indo- Krakatau (A.D. 1883; Simkin and Fiske, 1983; sidence causes large pressure changes, failure of nesia (e.g., Ambrose, 1998, 2003). These large Self, 1992). Exploration of the western Pacifi c the roof rocks, and collapse (e.g., Gudmunds- systems are commonly restless (e.g., Fig. 3C) has shown that caldera formation is not confi ned son, 1988; Lipman, 1997). The central depres- and thus pose a major monitoring challenge. to subaerial environments, but that silicic sub- sions often host thick sequences of caldera Another end member is represented by effu- marine calderas are also common in submarine fi ll. After caldera formation, these intracaldera sive eruptions of very viscous intermediate to arc volcanoes (e.g., Wright and Gamble, 1999; ignimbrites can be lifted up to form resurgent silicic magmas that create high-aspect-ratio lava Fiske et al., 2001; Tani et al., 2008). Moreover, domes at the center of the original depression fl ows, domes, and spines (Fig. 3D). Although silicic submarine calderas are often associated (e.g., Acocella et al., 2000). Eruptions of these effusive, these eruptions pose unique hazards re- with extensive Kuroko-type mineralization types are rare (Mason et al., 2004); thus recent lated to pressurization and collapse of dense lava (e.g., Iizasa et al., 1999). advances derive from new fi eld studies, analyti- plugs. Recent well-observed dome eruptions of Very large caldera-forming eruptions create cal techniques, analogue experiments, and nu- Mount St. Helens (1980–1986 and 2004–2008; inverse volcanoes, or central collapse depres- merical models rather than direct observation Swanson and Holcomb, 1990; Sherrod et al., sions surrounded by widespread pyroclastic of eruptive activity (e.g., Wilson and Hildreth, 2008); Unzen, Japan (1991–1995; Nakada et al., fans, such as the Taupo volcano, New Zealand 1997; Jellinek and de Paolo, 2003; Acocella, 1999), and Soufrière Hills volcano, Montserrat (Wilson, 1985). Large caldera systems are 2007; Cashman and Cas, 2011). Fascination (1995–present; e.g., Sparks and Young, 2002)

Geological Society of America Bulletin, May/June 2013 667 Cashman and Sparks have provided a wealth of observational data and coherent picture of well-studied volcanic decompression-induced crystallization (Annen about this eruption style, and show that growing systems (e.g., Blundy et al., 2008). Finally, in- et al., 2006). Sills form where magma moves domes (or cryptodomes) can generate (1) lateral tegration of geophysical information with nu- laterally; this can occur when rising magma volcanic blasts and sector collapse with debris merical models is providing new views into the encounters a rigidity or density barrier (Kava- avalanches, (2) subplinian and Vulcanian ex- evolution of magma storage regions (Pearse and nagh et al., 2006; Taisne and Jaupart, 2009), or plosive eruptions, and (3) (sometimes lethal) Fialko, 2010; Paulatto et al., 2012). Taken as a where the minimum principal stress is vertical. pyroclastic fl ows from collapse of active fl ow whole, the growing convergence of geophysical Sheeted sills probably constitute much of the fronts. A key factor for hazard assessment has and petrological evidence for the location and crust formed at mid-ocean ridges (e.g., Fialko, been the recognition of not only the longevity geometry of magma storage regions is encour- 2001). Dike-sill complexes may also amal ga- of some dome-forming activity but also the aging, particularly with regard to understanding mate to form magma reservoirs (e.g., Hildreth, rapid transitions from effusive to explosive ac- links between magma storage conditions and 2006; Menand, 2011; Fig. 4). Magma accumu- tivity that characterize eruptions of this type (e.g., volcanic eruptions. lation suffi cient for amalgamation requires heat Sparks, 1997). to be advected into the upper crust faster than Magma Chamber Formation heat can be conducted away or lost by hydro- DEVELOPMENT OF SHALLOW thermal circulation. Modeling by Annen (2009) MAGMA STORAGE SYSTEMS It is important to note that most magma never suggests that magma intrusion rates must ex- makes it to the surface. Estimates of the propor- ceed 10–3 km3/yr to allow magma chambers A critical control on eruption style is the pre- tion of intruded to extruded magma range from to form. eruptive history of shallow magma storage in 3 to 10 (e.g., Newhall and Dzurisin, 1989; White magma chambers. Magmatic systems are com- et al., 2006). From this perspective, intrusions Geophysical Evidence for Magma monly envisaged as interconnected crystal-melt can be viewed as failed eruptions (or, from an- Chambers mush zones and melt-dominated regions, or other perspective, eruptions may be viewed as magma chambers (Hildreth, 2006; Annen et al., failed intrusions). In either case, understand- Locations of active magma intrusion can be 2006), where the distinction between mush and ing the causes of, and interpreting the signs of, identifi ed from geophysical data and constrained magma is that of noneruptible and erupti ble magma arrest has become an important focus of by surface phenomena such as invigorated fuma- material. The threshold itself depends on fac- volcanological studies (e.g., Moran et al., 2011; roles or phreatic explosions. A particularly excit- tors like crystal size, shape, and strain rate (e.g., Bell and Kilburn, 2012). ing development in the past decade has been the Kerr and Lister, 1991; Cimarelli et al., 2011). Rising magma can stall for a variety of me- increasing use of interferometric synthetic aper- Below a critical melt fraction, the system can chanical reasons (Taisne and Tait, 2009). Dikes ture radar (InSAR) to identify and monitor in- be viewed rheologically as partially molten may not reach the surface if the driving pressure trusion sites (e.g., Pritchard and Simons, 2004; rock, while above the threshold, the system is is insuffi cient (e.g., Lister and Kerr, 1991), if the Biggs et al., 2009; Fournier et al., 2010; Riddick magma. The focus by the volcanological com- magma density is too high (Ryan, 1987), or be- and Schmidt, 2011). Regions of melt accumu- munity on (instantaneously) eruptible magma cause the volcanic edifi ce itself creates regions lation can also be imaged geophysically using has produced a different view of magmatic sys- of high stress below (Pinel and Jaupart, 2004). both active and passive seismic techniques. Pas- tems than that derived from those who study Dikes can suffer thermal death because of cool- sive source techniques, such as location of earth- (noneruptible) plutons, where the entire history ing (Taisne and Tait, 2011) or viscous death by quakes adjacent to magma bodies (Ryan, 1987), of the magmatic system is preserved. Our knowledge of magma storage systems has improved enormously over the past 25 yr thanks to developments in observational, ana- A B lytical, experimental, and numerical techniques. Geophysical techniques for imaging magma reservoirs have evolved from earthquake location (e.g., Scandone and Malone, 1985; Ryan, 1987) Intrusive to sophisticated tomographic and deforma- Complex tion inversion techniques (e.g., Koulakov et al., Mature magma 2011; Ofeigsson et al., 2011; Paulatto et al., chamber 2012). New analytical techniques for measuring the volatile contents of phenocryst-hosted melt inclusions (reviewed in Métrich and Wallace, 2008; Blundy and Cashman, 2008) and crystal zoning profi les (reviewed in Costa and Morgan, 2010) provide detailed constraints on magma Deep magma storage conditions that lead to volcanic erup- source tions. Additionally, experimental constraints on magma storage (reviewed in Picha vant et al., Figure 4. Proposed development of a magma chamber at Soufrière Hills volcano, Montserrat 2007) and the kinetics of phase transforma- (after Zellmer et al., 2003). (A) Repeated intrusion of magma from the deeper crust creates tions (reviewed in Hammer, 2008; Hamada a network of partially solidifi ed dikes and sills. (B) Continued intrusions reheat, remobi- et al., 2010) can be linked to geophysical and lize, and amalgamate melt pockets into a magma chamber; additional infl ux of deep mafi c geochemical observations to provide a detailed magma triggers an eruption.

668 Geological Society of America Bulletin, May/June 2013 How volcanoes work are the most common. Passive source seismol- Taken together, these studies show that mag- accumulation retards the ascent of new magma, ogy has improved dramatically through use of matic systems beneath many volcanoes and and thus extends the time of magma residence broadband and three-component seismometers; volcanic regions consist of localized magma in the upper crust prior to eruption. For this rea- new data have also raised new questions about chambers (zones of melt accumulation) concen- son, identical magmas rising beneath, or outside the source of volcano-related seismicity (e.g., trated at the top of much thicker regions of crystal- of, central magma storage regions may have the Neuberg et al., 2006; Harrington and Brodsky, melt mush. Importantly, large bodies of melt are same composition but different temperatures, 2007; Waite et al., 2008; Moran et al., 2008). At rarely detected, which suggests that most persis- crystallinities, and eruption styles (e.g., Frey and the same time, improvements to data inversion tent magma chambers are small and that the accu- Lange, 2011). Processes active within central techniques are providing increasingly detailed mulation of very large bodies of magma required (and open) magma chambers can be complex, three-dimensional views of active magmatic sys- to feed large ignimbrite eruptions is rare. This in including magma mingling/mixing, composi- tems (e.g., Chaput et al., 2012). turn suggests that melt is generated and stored as tional stratifi cation, disruption of cumulates, During the 1990s, geophysical experiments partial melt within the deeper crust for long peri- and assimilation of wall rock. Additionally, in mid-ocean-ridge environments showed that ods of time prior to being transferred rapidly to several recent studies document the presence of melt storage along fast-spreading ridges is typi- shallow levels (Annen et al., 2006). Support for diverse and intimately mixed individual crys- cally shallow (2–3 km) and confi ned primarily rapid magma transfer can be found in a recent tals with very different origins. These processes to thin along-axis sills (e.g., Singh et al., 2006). petrological study of magma accumulation prior are illustrated schematically in Figure 6, which A similar picture is emerging from recent activ- to the Bronze Age eruption of Santorini (Druitt shows both the magma storage and transport ity in Iceland, where deformation and seismic et al., 2012) and in a zircon chronology study of system of Shiveluch volcano, Kamchatka, and signals related to the 2010 eruption of Eyjaf- the Taupo volcanic system, New Zealand (Wilson the consequent diverse histories of individual jallajökull volcano, Iceland, suggest that this and Charlier, 2009). phenocrysts reconstructed from petrological eruption was fed by a complex network of sill- studies. Importantly, these observations require like magma bodies that may have extended to Petrologic Constraints on Magma Storage physical mechanisms of incorporating, and the base of the crust (e.g., Sigmundsson et al., and Eruption Triggers homoge niz ing, crystal cargo from very different 2010; Tarasewicz et al., 2012). In contrast, parts of the magma storage systems. How this magma beneath stratovolcanoes is apparently Petrology has long been used to study the occurs remains an important question. stored in narrow and vertically elongated re- evolution of magmatic systems. Over the past Constraints on the temperature, pressure, gions that may segregate into small melt pockets few decades, petrologic techniques have been volatile partial pressures, and oxidation state in (e.g., Lees, 1992; Waite and Moran, 2009; Pau- increasingly applied to questions of pre-erup- magma chambers can be obtained using geo- latto et al., 2012; Fig. 5). Upper-crustal magma tion magma storage, particularly those that thermometers, geobarometers, melt inclusion chambers beneath stratovolcanoes are probably may lead to eruptions. Here, we briefl y review studies, and comparison of natural mineral as- fed from larger magma accumulations at depth. insight about magma storage conditions that semblages with those produced in experiments An example is provided by Vesuvius, Italy. has been acquired from petrologic studies. We (reviewed in Pichavant et al., 2007; Blundy Here, a small shallow magma body (4–5 km then look at petrologic constraints on triggers of and Cashman, 2008; Putirka, 2008; Métrich depth) is underlain by a main magmatic system eruptive activity. and Wallace, 2008). Experimental studies of at 10–15 km depth, while the melt-bearing re- subduction-zone volcanoes suggest that magma gion extends to ≤30 km depth (Di Stefano and Petrologic Constraints on Magma Storage is commonly stored at 100–200 MPa under Chiarabba, 2002). More generally, zones of volatile-saturated conditions. These experi- plate convergence often show midcrustal anom- Once a magma chamber has formed, it de- ments also show that exsolution of H2O accom- alies that may be associated with melt accumu- termines both the nature of magmas that can panying magma ascent changes the stability of lation (Brown et al., 1996; Zandt et al., 2003). erupt and the fate of new melt inputs. Magma some crystal phases, most notably plagioclase. Thus, the preserved phase assemblage (phases, phase proportions, and phase compositions) WEoften provides tight constraints on storage con- 0 ditions just prior to eruption (Fig. 7A). Informa- Low velocity tion on pre-eruptive magma storage can also magma be derived from analysis of phenocryst-hosted Figure 5. Schematic illustration melt inclusions. The preserved volatile content of the magma storage system 5 of melt inclusions, in particular, can be used to infer the pressure (depth) of crystal formation beneath Mount St. Helens show- High velocity and thus the nature of magma storage systems ing earthquake locations (open plug circles) in country rock (high ve- characteristic of different volcano types (Fig. locity) surrounding earthquake- 10 7B). Early-erupted samples from large (ignim- free zones of magma storage Low velocity brite-producing) rhyolitic eruptions typically Depth (km) (low velocity). Modified from magma preserve melt inclusions that are H2O-rich but Lees (1992). have lost much of their CO2, suggesting pro- 15 ? tracted storage at (often) 150–200 MPa (open Feeder circles, Fig. 7B). Later-erupted samples from the same eruptions have melt inclusions that 5 km ? are more enriched in CO2, and often encompass 20 a wider range in pressure than early-erupted

Geological Society of America Bulletin, May/June 2013 669 Cashman and Sparks

time scales of months to decades (e.g., Hawkes- worth et al., 2004; Costa and Morgan, 2010). ASCENT AND VOLATILE EXSOLUTION DRIVES Similar time scales are recorded by U-series CRYSTALLIZATION studies of magma degassing (e.g., Condomines et al., 2003; Berlo et al., 2004).

Triggering Eruptions

Two end-member models have emerged for eruption triggering. First, there is clear evidence CRYSTALS RECORD CHANGES IN fO2 that melt recharge events may trigger eruptions PLAGIOCLASE RECORD by mobilizing stored and partially crystalline PRESSURE OSCILLATIONS AT TOP OF CHAMBER magma (e.g., Murphy et al., 2000; Ruprecht and Cooper, 2012). Second, evolved melt may be rapidly and effi ciently extracted from mush zones and transported to shallow storage regions. These two mechanisms are not mutually exclusive, and they may act in tandem to fuel some eruptions. Other triggering mechanisms include buildup of SMALL LATE-STAGE pressure in crystallizing, water-supersaturated HBL & PLAG INPUTS OF OL & OPX RESORB magma (Tait et al., 1989) and tectonic triggering HOT MAGMA CRYSTALLIZE DURING QUENCH & DISPERSE (e.g., Gravley et al., 2007). ASCENT Mafi c recharge can trigger eruptions when LARGE LATE- hot, low-viscosity, crystal-poor melt batches STAGE BASALT interact with cooler stored magma that is typi- INTRUSION EARLY-FORMED CUMULATES cally more evolved and more crystalline. The (semirigid) mush zone serves to both stabilize the magma and to act as a trap, or rheological INPUTS OF UNDERSATURATED barrier, to eruption of either the evolved crystal- HYDROUS MAGMA line melt or the low-viscosity recharge magma (e.g., Kent et al., 2010). The recharge magma Figure 6. Schematic diagram showing reconstruction of the processes may disrupt the crystal network by fl uxing gases taking place in the open-system magma chamber of Shiveluch vol- (Bachmann and Bergantz, 2006), by heating to cano, Kamchatka, before and during eruption, based on petrologi- cause convective self-mixing (e.g., Couch et al., cal studies. Phenocryst zoning patterns provide evidence of multiple 2001), or by fracturing (Wright et al., 2011). Al- episodes of magma intrusion, crystallization, and resorption at though the recharge magma is commonly mafi c, varying pressures and oxygen fugacities, as well as late-stage crys- cryptic recharge of (hotter and less viscous) tallization driven by magma ascent and volatile exsolution. Figure is silicic magma may also serve as an eruptive after Humphreys et al. (2006). trigger (e.g., Smith et al., 2009). There is also growing evidence of effi cient segregation and upward migration of rhyolitic melt samples (fi lled circles in Fig. 7B), suggesting 2003), and arrival of recharge magma into the from midcrustal “mush” zones. First suggested complex patterns of magma withdrawal dur- shallow system (Murphy et al., 2000). by Eichelberger et al. (2000), rhyolite melt seg- ing later stages of caldera-forming eruptions. In The past decade has also seen rapid improve- regation has been particularly well documented contrast, open-system andesitic volcanoes, such ments in microanalytical techniques that are by zircon dating of the products of Taupo vol- as Popocatepetl, Mexico, preserve phenocryst- providing new insight into the details of magma cano, New Zealand. Here, melt segregation ap- hosted melt inclusions that suggest that magma storage conditions. Time scales of magmatic pears to have been both effi cient and rapid, such was supplied from a large pressure range despite activity are constrained by isotopic (particu- that large (~500 km3) volumes of melt may have small eruptive volumes (fi lled diamonds in Fig. larly U-series isotopes) and diffusion studies. accumulated in hundreds to a few thousands of 7B). These data can be correlated with seismic U-series studies place constraints on phenocryst years (e.g., Charlier et al., 2005). Rapid melt seg- and gas geochemistry data when the samples are residence times within magma storage regions regation and shallow accumulation of rhyolitic well constrained (i.e., both timing and eruption (e.g., Zellmer et al., 2003; Cooper and Reid, melt prior to large caldera-forming eruptions conditions are known; e.g., Blundy et al., 2008; 2008). These studies refl ect the integrated age of are consistent with melt inclusion evidence for Saunders et al., 2012). Additionally, discrepan- the crystal population and commonly yield time sill-like geometries (e.g., Blundy and Cashman, cies between petrologic and geophysical moni- scales of thousands of years. In contrast, diffu- 2008), fi eld and theoretical evidence for the in- toring data may provide critical information on sion studies take advantage of chemical zon- herent instability of such melt accumulations late-stage processes (those that occur just prior ing profi les in phenocrysts caused by magma (e.g., Jellinek and DePaolo, 2003; De Silva et al., to eruption) such as magma rise (Blundy and recharge or depressurization. This approach is 2006), thermal models (Annen, 2009), and lack Cashman, 2001), fl ushing of volatiles through more likely to refl ect events responsible for trig- of geophysical evidence for large melt accumu- the magma chamber (Hammer and Rutherford, gering eruptive activity, and commonly yields lations in most magmatic settings.

670 Geological Society of America Bulletin, May/June 2013 How volcanoes work

months, or even years. For example, about eight A 0 0 B 1600 Trid 250 weeks of precursory activity (seismicity, defor- Qz 2 mation, and explosions) preceded climactic Amph-out 30 20 1200 eruptions of both the 1980 eruption of Mount 100 10 4

% crystallization Depth (km) St. Helens, United States, and the 1991 erup- 6 150 tion of Pinatubo volcano, Philippines. Both of Qz-in these precursory sequences involved movement 200 8 800 Water saturation 2 of magma to shallow levels, as evidenced by

10 CO (ppm) the growth of a cryptodome (Mount St. Helens)

Liquidus 300 400 or dome (Pinatubo). However, a time delay of Pressure (MPa) 12 50 hours (Mount St. Helens) to a few days (Pina- 14 tubo) between eruption initiation and full devel- 400 0 opment of climactic activity suggests that the 700 800 900 1000 0 1 234567conduit did not develop full connectivity until H O (wt%) Temperature (°C) 2 after the main eruptive phase had commenced Figure 7. Magma storage conditions determined from (A) phase equilibria experiments and (Scandone et al., 2007). (B) phenocryst-hosted melt inclusions. Phase equilibria data are for magma erupted from Once formed, conduit geometry can evolve Mount St. Helens in 1980; solid square shows location of last magma storage as determined in space and time as the result of mechanical from the phenocryst assemblage and Fe-Ti–oxide thermometry (modifi ed from Blundy and and/or thermal erosion, magma solidifi cation, Cashman, 2001). Melt inclusion data are from early (open circles) and late (fi lled circles) and changing stress conditions. Conduit evolu- phases of the Bishop Tuff (Wallace et al., 1999) and Oruanui (Liu et al., 2006) ignimbrites tion is particularly apparent in basaltic erup- and Popocatepetl stratovolcano (solid diamonds; Atlas et al., 2006). Note that early phases tions, which typically initiate with a “curtain of fi re” as magma-transporting dikes intersect the of ignimbrite eruptions have melt inclusions that are H2O-rich and CO2-poor, suggesting long residence times within upper crust. In contrast, both late-stage ignimbrites and open- surface, but rapidly focus into one (or a few) localized vents. Processes that promote focusing system volcano Popocatepetl preserve melt inclusions with a wide range in CO2 (pressure) in these systems most likely involve feedbacks for a limited range of H2O. Figure is modifi ed from Blundy and Cashman (2008). among fl ow, magma rheology, and cooling (Bruce and Huppert, 1989). For example, ad- VOLCANIC CONDUITS is commonly assumed that magma ascends via vection of heat by rapid fl ow through the widest dike propagation, at least until it reaches shal- part of the initial dike will slow, or eventually Prior to the mid-1980s, the vigor of erup- low levels. The speed of dike propagation de- reverse, rates of chilled margin growth (Holness tive activity was linked directly to the extent to pends principally on magma viscosity (Kerr and and Humphreys, 2003); at the same time, lower which stored magma was saturated, or under- Lister, 1991) and can be fast for basaltic magmas fl ow rates through narrower regions will pro- saturated, with volatile components. The 1990s (decimeters to meters per second; e.g., Linde mote cooling and eventual solidifi cation of dike saw a change in emphasis to conditions of et al., 1993). Dike width is controlled by magma extremities. Viscosity changes caused by degas- magma transport from magma storage regions pressure through elastic deformation of the wall sing and crystallization should produce a similar to the surface (that is, the role of conduits). This rock, with magma fl ow rate proportional to the fl ow focusing if the rheological changes occur shift was largely the result of detailed observa- product of the cube of the dike width and the preferentially at the dike margins (for example, tions of effusive eruptions of Mount St. Helens, length. For this reason, dike-fed eruptions show in regions of high shear). United States, and Unzen volcano, Japan, cou- strong interactions between eruption rate and The geometry of volcanic conduits can also pled with recognition of the extent to which the magma pressure (e.g., Costa et al., 2006). Dike evolve by mechanical processes. Mechanical physical properties of magma could change dur- closure may cause eruptions to end (or vents to erosion is most likely where dikes change ori- ing transport because of decompression-driven shift) if the pressure becomes too low to drive entation, or at shallow levels in explosive vents. volatile exsolution (e.g., Dingwell et al., 1996) continued magma fl ow or if magma cools and Exposures in caldera walls show that dikes are and crystallization (e.g., Cashman, 1992). These solidifi es on the dike walls. These scenarios can commonly segmented; offsets or jogs between rheological changes set up complex feedbacks be distinguished if there are good temporal con- segments are regions of complex brittle defor- between conditions of magma ascent and result- straints on mass fl ux. For example, documenta- mation, brecciation, and dilation that can local- ing styles of eruptive behavior (e.g., Melnik and tion of a linear decrease in magma supply to the ize fl ow. Xenoliths generated by deformation Sparks, 1999). Kūpaianaha vent of Kilauea volcano, Hawaii, associated with localization can be removed between April and November of 1991 provided by fl owing magma and transported to the sur- Conduit Construction and Evolution evidence of a gradual loss of driving pressure face (Brown et al., 2007; Kavanagh and Sparks, with time (Kauahikaua et al., 1996). Alterna- 2011). Additionally, protracted effusive erup- With the exception of open-system volcanoes tively, dikes may fail to close completely when tions may create complex conduit systems. An such as Stromboli, Italy, and Villarica, Chile, cooling at the thin edges is combined with in- unusual opportunity to view such a system in a magma storage regions are not connected to elastic deformation (Daniels et al., 2012). recently active volcanic conduit was provided the surface; it is this isolation that allows them Eruptions of viscous silicic magma are by the Unzen (Japan) drilling project. Drilling to develop suffi cient overpressure to generate also dike-fed (e.g., Mastin and Pollard, 1988; of the conduit system that fed a 1991–1995 eruptive activity. Thus, magma must construct Roman and Cashman, 2006), although silicic dome-building eruption revealed a wide (500 m) a pathway (conduit) to the surface. Conduit dikes ascend suffi ciently slowly that arrival conduit zone consisting of numerous individual construction is not well understood, although it of magma at Earth’s surface may take weeks, feeder dikes (e.g., Sakuma et al., 2008).

Geological Society of America Bulletin, May/June 2013 671 Cashman and Sparks

Explosive eruptions also create conditions 0.1 of rapid (syneruptive) crystallization and associ- that promote mechanical erosion. In particular, 8.5 MPa/s ated heterogeneous bubble nucleation. cylindrical near-surface conduits develop when equilibrium solubility0.025 MPa/s The number and proportion of crystal phases early magma is either sufﬁ ciently overpressured also provide information on conditions of to excavate a conduit or sufﬁ ciently underpres- 50 0.003 MPa/s magma ascent, particularly when the stability

sured to cause wall rocks to fail, fall into the HOMOGENEOUS NUCLEATION of the crystal is controlled by the water con- conduit, and be transported out of the conduit by HETEROGENEOUS tent of the melt (e.g., Hammer et al., 2000; Tora-

high-speed explosive fl ows (Sparks et al., 2007; ∆ maru et al., 2008; Blundy and Cashman, 2008). Barnett and Lorig, 2007). In powerful explosive P ~ 150 MPa The time required for crystals to nucleate and eruptions, the level of fragmentation, and thus 100 grow in response to water exsolution depends the depth of the resulting cylindrical conduits, on the melt composition and temperature (dif- can extend to kilometers. Early phreatic or phre- fusion rate), the undercooling (driving force) for atomagmatic stages of eruptive activity may also Pressure (MPa) crystallization, and the presence or absence of form cylindrical near-surface pathways that can 150 phenocryst phases. For a single melt composi- then be used by magma fed from a deeper dike. tion, decompression experiments confi rm that both the number density and volumetric pro- Synascent Changes in Magma Properties portion of plagioclase crystals record the con- 200 ditions of decompression (reviewed in Blundy As magma ascends toward Earth’s surface, 024 68and Cashman, 2008; Hammer, 2008). Experi- decompression causes some volatile phases to Dissolved water (wt%) ments also confi rm observational evidence that exsolve and some solid phases to precipitate. decompression-induced crystallization can occur These phase transformations affect the density Figure 8. Degassing/vesiculation paths for on eruptive time scales (e.g., Geschwind and and rheology of the magma and, to a lesser rhyolite decompression experiments per- Rutherford, 1995; Hammer and Rutherford, extent, its temperature. The past few decades formed at conditions that promote both 2002). The combined effects of vesiculation and have seen extensive research on the kinetics of homogeneous and heterogeneous bubble crystallization during magma ascent have pro- the phase transitions, the rheology of complex nucleation. Heterogeneous nucleation per- found consequences for the rheological evolu- (multiphase) suspensions, and the evolution of mits the melt to maintain equilibrium de- tion of ascending magma, and for the course of the gas phase, all of which are important for gassing paths. In contrast, homogeneous vol canic eruptions. understanding the highly nonlinear dynamics nucleation occurs only at large supersatu- of conduit fl ow processes that control eruptions. rations (ΔP) regardless of magma ascent Magma Rheology rates; once nucleation occurs, ascent rate Volatiles, Bubbles, and Crystals controls the degassing path. Figure is modi- Rheology refers to fl ow behavior (that is, fi ed from Mangan and Sisson (2000). the deformational response to imposed stress). Volatiles are more soluble in silicate melts Magma rheology is usually measured by at high pressure than at low pressure (e.g., magma viscosity, which varies depending on

Newman and Lowenstern, 2002; Papale et al., exsolution of a mixed (H2O-CO2) volatile phase the melt composition and temperature as well 2006). For this reason, decompression of (e.g., Cashman, 2004). This conclusion has im- as the bubble and crystal content. The rheologi- volatile-saturated melt causes exsolution of the portant implications for conditions of magma cal properties of magma govern the dynamics of volatile phase as bubbles (vesiculation). The fragmentation in silicic eruptions where delayed magma chambers, the wall friction generated by rate of vesiculation is controlled by the rate nucleation may generate explosive vesiculation conduits fl ows, and therefore the rate of erup- of bubble nucleation and growth, which de- bursts at high overpressures (e.g., Mangan et al., tion, the kinetics of crystal and bubble forma- pends not only on the degree of supersatura- 2004; Scandone et al., 2007). Additionally, these tion, and the fl ow of lava on Earth’s surface. tion caused by the decompression but also on data show that the kinetics of bubble and crys- Critical constraints on the rheology of silicic the surface tension and viscosity of the melt tal formation are intimately linked and together melts have been provided by experimentally phase (e.g., Mangan and Sisson, 2005) and the may control transitions in eruption style. Unfor- calibrated models for melt viscosity as a func- availability of nucleation sites (e.g., Hurwitz tunately, there are no equivalent data for bubble tion of composition, temperature, and water and Navon, 1994). Results from decompres- nucleation in mafi c melts because of experimen- content (e.g., Dingwell, 1998; Giordano et al., sion experiments show that in rhyolitic melts, tal challenges. Measurement of bubble number 2008). Most silicate melts are Newtonian (for homogeneous nucleation (that is, nucleation densities in the pyroclastic products of low- to details, see Giordano and Dingwell, 2003) and within the melt) has to overcome substantial moderate-intensity eruptions, however, suggests have viscosities that vary from less than 0.1 Pa s energy barriers and therefore requires very that there is little to no activation energy barrier to over 1012 Pa s. Silicate melts that are either al- large overpressures (100–150 MPa; Fig. 8). In for bubble nucleation in mafi c melts (e.g., Rust kalic or hydrous, however, show non-Arrhenian contrast, the undercooling required for bubble and Cashman, 2011), although textural studies behavior controlled by the effect of alkalis/water nucleation is much lower if nucleation can oc- of mafi c Plinian deposits show that these sys- on the melt structure. Understanding the effect cur heterogeneously on crystal surfaces (Hur- tems can attain bubble number densities that ap- of water, in particular, is important for modeling witz and Navon, 1994; Gardner et al., 1999). proach those of silicic pumice (e.g., Sable et al., the behavior of ascending and degassing mag- Bubble number densities preserved in pyro- 2006; Costantini et al., 2010; Vinkler et al., mas. Importantly, the glass transition also varies clastic material (pumice) produced by silicic 2012). Very high bubble number densities may as a function of melt composition, temperature, Plinian eruptions suggest that bubble nucleation refl ect large supersaturations generated by rapid and shear rate (e.g., Dingwell and Webb, 1989; is commonly homogeneous and controlled by magma decompression, or possibly the effects Webb and Dingwell, 1990). The glass transition

672 Geological Society of America Bulletin, May/June 2013 How volcanoes work is a kinetic barrier (relaxation time scale) that crystal volume fraction for particle interactions average shear rate and shear stress experienced separates liquid from glassy behavior, which in depends on crystal shape, with higher aspect by a sample can be determined from the dimen- turn determines the behavior of silicate liquids ratios allowing interaction at lower volume sions of moderately deformed bubbles (Rust when strained. One application of these data has fractions (Saar et al., 2001; Walsh and Saar, et al., 2003). Suspensions of bubbles in viscous been to defi ne magma fragmentation in melt- 2008). For this reason, crystallization caused liquids are shear-thinning (Fig. 9C): Addition of viscosity shear space (e.g., Gonnermann and by volatile exsolution during rapid decompres- bubbles increases the viscosity at low Ca (small Manga, 2003). An important outcome of such an sion, which generates numerous small elongate bubbles, low strain rates) but decreases viscos- analysis is to demonstrate that viscous rhyolitic or platy plagioclase crystals, can cause the ap- ity at high Ca (large bubbles, high strain rates). glass, in particular, may repeatedly break and parent viscosity to change by many orders of Bubble suspensions are more strongly shear- φ re-anneal during slow ascent at shallow levels . magnitude as magma ascends from the reservoir thinning at higher bubble volume fractions ( b). Repeated fracture, in turn, may explain the char- to the surface (e.g., Sparks, 1997). The effect Under these conditions, the relative viscosity φ Ca acteristic “hybrid” earthquakes that accompany of crystals on viscosity is enhanced by the ten- (µsusp/µmelt) approaches 1 – b at high , because extrusion of silicic domes (Tuffen et al., 2003; dency of the remaining melt to become more si- the bubbles are suffi ciently deformed that resis- Neuberg et al., 2006). licic (more viscous) as crystallization proceeds tance to fl ow is provided only by the melt frac- φ When the melt contains suspended crys- (e.g., Cashman and Blundy, 2000). tion (= 1 – b). Shear-thinning behavior means tals and bubbles, the magma can develop non- The effect of bubbles on magma rheology de- that ascent of bubbly viscous magma through Newtonian rheological properties. The addition pends on both bubble size and shear rate (e.g., volcanic conduits probably occurs by plug fl ow, of a small volume fraction of crystals causes an Rust and Manga, 2002; Pal, 2003; Llewellin with localization of shear along the conduit mar- increase in magma viscosity; when the crystal and Manga, 2005). The deformation behavior gins (e.g., Llewellin and Manga, 2005). content is suffi ciently high, crystal interactions of bubbles in a shear fl ow is described by a generate a yield strength (e.g., Lejeune and nondimensional parameter called the capillary Conduit Controls on Eruption Style Richet, 1995; Mueller et al., 2010; Castruccio rγμ· number (Ca = , where r is bubble radius, γ· et al., 2010; Fig. 9A) and either shear-thicken- Γ The fl ow of magma along conduits is driven ing (shear-dilatancy) or shear-thinning behavior is strain rate, µ is melt viscosity, and Γ is sur- by pressure gradients between the deep source (e.g., Costa et al., 2009; Fig. 9B). The critical face tension). In dilute bubble suspensions, the and the surface, and opposed by both friction

12 Bingham Shear-thinning Newtonian σy

10 Stress viscosity (Pa s) 8 10

Log Shear rate 6 Homogeneous 40 vol% 60 vol% Polycrystal C 10 liquid (1) (2) (3) (4) (5) µr = 1 for all φ at Ca = 0.645 Crystal Content Newtonian non-Newtonian 1 φ = 00.1.1 Flow Flow φ = 00.3.3

φ = 00.5.5 Relative viscosity µr

0.1 0.01 0.1 110 100 1000 (1) (2) (3) (4) (5) Capillary number Ca

Figure 9. Basics of magma rheology. (A) Effect of adding spherical particles; y-axis shows viscosity normalized to reference (melt) viscosity; dashed lines at 40% and 60% particles show locations of rapid viscosity increase and maximum packing, respectively. Diagrams show schematic representation of particle concentrations (modifi ed from Lejeune and Richet, 1995). (B) Schematic stress–shear rate diagram illustrating σ different rheologies; y is yield strength (minimum stress that must be overcome for fl uid deformation). (C) Effect of adding bubbles; viscosity is normalized to bubble-free values and shown as a function of Ca for different bubble concentrations (φ; modifi ed from Pal, 2003).

Geological Society of America Bulletin, May/June 2013 673 Cashman and Sparks along the conduit margins and the tendency of dome extrusion or destruction of an impermea- infl ux of new magma. Bubbles can have a pro- magma to degas and solidify. The rate of magma ble magma plug when overpressure exceeds a found effect because gas is highly compressi- ascent controls the eruption style (explosive or threshold (e.g., Wylie et al., 1999; Druitt et al., ble; for this reason, bubble-bearing magma can effusive) by modulating the extent to which 2002b), or (2) crystallization coupled with sustain much higher chamber pressures dur- exsolving gas is retained within, or lost from, rheological changes (Melnik and Sparks, 1999, ing eruption, and therefore erupt much more magma during ascent. Gas loss, in turn, is con- 2005). Intermediate time scales are marked by magma, than a chamber without bubbles (Hup- trolled by the relative rates of bubble rise, bubble 6–7 wk cycles of earthquakes, tilt, and eruptive pert and Woods, 2002). For this reason, the rise coalescence, and the development of permeable activity that may be explained by opening and and accumulation of exsolved gas at the top of a pathways in magma and surrounding host rocks. closing of dikes because of pressure fl uctuations magma chamber (a consequence of crystalliza- As bubble rise is controlled primarily by magma (Roman et al., 2006). The longest time scale tion of volatile-saturated magma) can generate viscosity, the orders-of-magnitude variation in is represented by 2–3 yr alternating periods large pressure increases (Tait et al., 1989), as viscosity between mafi c and silicic melts means of dome growth and quiescence. Long-time- can infl ux of new magma. that mechanisms of gas loss are very different in scale behavior can be explained if the magma mafi c and silicic magmas. chamber and conduit act like capacitors, that Gas Behavior during Magma Ascent is, if they store energy because of elastic de- Modeling Conduit Flow formation of the wall rocks and then discharge Once mobilized, magma ascends because of magma episodically when the pressure exceeds volatile exsolution; ascent is therefore modu- Modeling the fl ow of magma through volcanic some threshold. The time scale for this process lated by conditions of both vesiculation and gas conduits requires coupling equations of mass refl ects the elastic relaxation of the chamber, escape, which depend critically on the viscosity and momentum with expressions for changes with longer periods being the consequence of of the melt. The low viscosity of basaltic melts in phase proportions, and resulting changes in larger chamber volumes (Barmin et al., 2002). allows bubbles to separate from the ascending rheology. The numerous interacting factors that Periodic behavior may occur when magma vis- magma when the rise rate of individual bubbles control fl ow rates explain the very rich variety cosity increases during magma ascent (because (the “drift velocity”) is rapid relative to the of volcano behaviors (reviewed in Melnik et al., of devolatilization ± crystallization) given an rate of magma ascent (Fig. 11A). As the ratio 2008). These complex interactions can be illus- appropriate input fl ux and input/output viscos- between drift velocity and ascent velocity in- trated by a simplifi ed reference case for effusive ity ratio (e.g., Melnik et al., 2008; Fig. 10A). creases, the distribution of the gas phase in the eruption from a pressurized magma chamber Magma rheology, particularly yield strength, liquid changes from distributed bubbles (bubbly with elastic walls. Under these conditions, the will affect the overpressure threshold that must fl ow) to large conduit-fi lling bubbles (slug fl ow) fl ow rate of magma through a cylindrical con- be exceeded for magma ascent (Fig. 10B). to a continuous gas phase concentrated in the duit or parallel-sided fracture is controlled by The reference case described here includes center of the conduit (annular fl ow). the pressure gradient and wall friction, which, in numerous simplifi cations. For example, in the Two-phase fl ow regimes are well character- turn, refl ect both conduit geometry and magma reference case, the magma chamber pressure ized for water-gas systems (e.g., Mudde, 2005). viscosity. If magma viscosity and conduit di- changes only with volume, whereas chamber Only recently, however, have volcanologists mensions are constant, the magma discharge rate pressure may actually depend on other variables attempted to scale experiments and develop will decline exponentially with time as pressure such the presence of bubbles and crystals or the numerical models to examine two-phase fl ow and volume in the chamber decline (e.g., Stasiuk et al., 1993). Deviations from this simple model will occur with (1) the growth of a lava dome, AB3 which can increase the column weight; (2) formation of chilled margins, which can reduce conduit width; (3) elastic deformation of the dike 3 20 itself; (4) viscous dissipation at the fl ow margins where shear rates are high; and (5) variations of RAPID ASCENT magma composition, temperature, and gas con- constant viscosity tent, which can change viscosity. Important feedbacks develop during magma 4 10 NEWTONIAN BINGHAM ascent because of the competing effects of Discharge rate

PERIODIC BEHAVIOR Discharge rate (m /s) buoyancy (vesiculation and gas loss), viscosity σ 2 y = 0.05 σ (which changes with volatile loss and crystal y = 0.1 σ 1 y = 0.2 formation), and wall friction. Such feedbacks Q(IN) > Q(OUT) 0 may explain, for example, three different time Chamber pressure 130 140 150 scales of episodic behavior that have been iden- Chamber pressure (MPa) tiﬁ ed at Soufriere Hills volcano, Montserrat (e.g., Costa et al., 2006; Wadge et al., 2008). Figure 10. (A) General steady-state solution of feedbacks among magma input, overpres- The shortest time scale of several hours to a surization, and eruption. When decompression-related changes in magma viscosity are lim- few days is recorded in deformation and seis- ited, system proceeds as 1–2-3–4; periodic behavior is most likely when magma input lies mic data, and in patterns of dome extrusion and in gray shaded region. (B) Comparison of steady-state solutions for magmas of Newtonian σ Vulcanian explosions (e.g., Voight et al., 1999). and Bingham rheology (yield strength y in MPa). Bingham rheology prohibits discharge This time scale has been explained by (1) gas between eruptions and produces higher chamber pressures prior to renewal of eruptive pressure cycles that generate either stick-slip activity. Figure is modiﬁ ed from Melnik et al. (2008).

674 Geological Society of America Bulletin, May/June 2013 How volcanoes work

Figure 11. (A) Bubble rise A B time (drift velocity) as a func- 4 BUBBLES DEFORM tion of bubble radius and melt MSR = 1000 m 3/s GAS ESCAPES viscosity. Shaded area shows 0 magma ascent rates for typical Pm magma supply rates (MSR), MSR = 0.1 m3 /s 8 calculated assuming a cylindri- –4 4 6 MAGMA cal conduit of radius = 10 m, DENSIFIES and shows that drift velocity is –8 BASALT (100 Pa s) RHYOLITE (10 Pa s) most likely to exceed ascent ve- ANDESITEDACITE (10 Pa (10 s) Pa s) locity in low-viscosity (basaltic) BUBBLES CONNECT Magma ascent velocities magma. (B) Illustration of hys- –12 calculated assuming a LIMIT VESICULARITY Increasing permeability teresis in permeability-porosity Log bubble rise rate (m/s) cylindrical conduit of Pc 10m diameter data. Initial expansion occurs ISOLATED BUBBLES EXPAND in isolated bubbles until perco- –16 –3 –11 3 Increasing vesicularity lation threshold (Pc) is reached Log bubble radius (mm) at ~60% vesicularity. Perme- ability then increases rapidly to a permeability threshold (Pm) that limits further expansion. Deformation of vesicular magma permits gas loss (porosity reduction) without loss of permeability as bubbles deform, until bubble collapse reduces permeability (bubbles again become isolated). Figure is modifi ed from Rust and Cashman (2004). regimes in viscous fl uids and large conduits. tion threshold) will form at volume fractions as foam) is enhanced when permeability-porosity When gas is introduced into static liquid col- low as 30% (Sahimi, 1994). However, evidence curves are hysteretic, such that high permeabili- umns from below, high-viscosity fl uids enhance from both pumice samples and recent experi- ties are maintained as bubbles collapse, thereby bubble coalescence by decreasing the drift ve- ments suggests that bubbles in rapidly vesicu- facilitating gas escape (Fig. 11B). Horizontal gas locity of individual bubbles, thereby stabilizing lating magmas do not always coalesce (become fl ow is enhanced at low pressures when fractured slug fl ow at the expense of bubbly fl ow (e.g., connected) at low vesicularities, and instead wall rocks are at hydrostatic, or atmospheric, James et al., 2009; Pioli et al., 2012). Internal attain a connectivity, or percolation, threshold pressure. Porosity decrease in the upper parts vesiculation, the most likely source of distrib- (Pc) at vesicularities of ~60%–70% (Rust and of the conduit can also occur if vertical gas fl ow uted bubbles in volcanic systems, has not been Cashman, 2011). The permeability threshold exceeds the ascent rate of host magma (Melnik studied experimentally from the perspective (Pm) for suffi ciently rapid gas escape to pre- and Sparks, 1999). In this scenario, hysteretic of two-phase fl ow regimes. Numerical models vent continued magma expansion is probably permea bility functions predict rapid formation suggest that large conduit-fi lling bubbles may slightly higher than the percolation threshold of compaction waves manifested by alternating be dynamically unstable during buoyancy- (Fig. 11B). Available experimental data suggest regions of high and low porosity (Michaut et al., driven ascent (Suckale et al., 2010) and that that the percolation and permeability thresholds 2009). Degassing-driven crystallization may fur- cyclic patterns of fl ow developed in two-phase may increase with increasing melt viscosity, and ther enhance the hysteresis of porosity-permea- bubbly magmas may explain the strong pulsing decrease with increasing sample crystallinity, bility relationships in viscous magmas, as shown of Hawaiian , Strombolian, and violent Strombo- although these hypotheses need to be tested by by the maintenance of high permeabilities to lian activity (e.g., Manga, 1996; Slezin, 2003). If additional experiments. low bulk vesicularities in crystal-rich andesites crystals are present, gas migration may be hin- The high viscosity of silicic melts also pro- (Melnik and Sparks, 2002). dered if bubbles are trapped within the crystal motes bubble deformation (by increasing Ca). network, or aided by increased bubble coales- Evidence for bubble deformation can be found Controls on Fragmentation cence within melt pathways. Either case will af- in (1) the prevalence of tube (elongated bubble) fect fl ow regimes (Belien et al., 2010). Together, pumice in high-intensity silicic eruptions (e.g., If gas is retained within magma rather than these studies suggest that simple two-phase fl ow Wright et al., 2006); (2) observed bubble fl at- lost to wall rocks or the atmosphere, then as- interpretations of mafi c eruptive activity should tening; and (3) the common occurrence of pyro- cending magma will erupt explosively. Rapid be reconsidered. clastic obsidian in subplinian eruptions, which vesiculation (and expansion) under closed-sys- Bubble rise is suffi ciently hindered in vis- probably forms by effi cient gas loss along con- tem conditions accelerates magma to the sur- cous magmas that bubbles will remain in the duit walls (Rust and Cashman, 2007) and may face, as illustrated by the popular Mentos® and melt from which they formed unless they con- record shear-enhanced permeability develop- Diet Coke® experiments (Coffey, 2008). These nect to create permeable networks. Gas escape ment (e.g., Okumura et al., 2009). two processes—expansion and acceleration— through a permeable foam may not only allow Gas escape through the permeable magma form the core of fragmentation theory. In vol- degassed lavas to form from originally gas-rich leads to some interesting nonlinear dynamics. canol ogy, “fragmentation” denotes the transition magma (Eichelberger et al., 1986), but may If bubbles within the magma are suffi ciently from a melt (± crystals) with included bubbles also explain sharp transitions between explo- connected to supply gas to the wall rock, rapid to a continuous gas phase with suspended drop- sive and effusive styles of activity (e.g., Jaupart horizontal gas escape can be driven by pressure lets or particles (Fig. 12). Fragmentation may and Allegre, 1991). Permeability is commonly differences between the magma and low-pressure be ductile or brittle; in general, fragmentation modeled using percolation theory, which shows (wall rock) environments (Jaupart and Allegre, is ductile in low-viscosity (basaltic) melts and that a touching network of spheres (the percola- 1991). Gas escape (and collapse of the magma brittle in high-viscosity (silicic) melts.

Geological Society of America Bulletin, May/June 2013 675 Cashman and Sparks

Figure 13. Cumulative total Hawaiian ﬁre fountains grain-size distributions (TGSDs) Bubble size for different eruption styles. Plinian pumice } 100

Steeper curves represent better Strombo

sorting. Available data suggest Hawaiian a systematic increase in median 80 violent Strombolian

l Vulcanian ian fragmentation grain size from Plinian-sub- Plinian-subp plinian to Vulcanian–violent 60 Strombolian to Strombolian- Hawaiian eruption styles, al- 40 linian though more data are needed.

Also shown is range of bubble % Deposit mass > d 20 sizes observed in individual pyroclasts from Plinian and 0 saturation exsolution Hawaiian eruptions. Overlap –6 –5 –4 –3 –2 –1 0 in bubble size distributions Log d, grain diameter (m) (BSDs) and TGSDs for Plinian eruptions suggests a direct relationship between vesiculation kinetics and fragmentation. Figure 12. Schematic representation of vesic- Figure is modifi ed from Rust and Cashman (2011). ulation and fragmentation within a volcanic conduit. Saturation level requires saturation with at least one volatile phase; although pumice clasts. The lower values derive from porosity (permeability; Spieler et al., 2004; shown here as top of magma chamber, CO2- minimum preserved vesicularities and assume Koyaguchi et al., 2008; Mueller et al., 2008) rich magmatic systems may be volatile-satu- that higher vesicularities record postfragmenta- and fragmentation effi ciency that is controlled rated throughout the upper-crustal storage tion expansion prior to quenching. The critical by the applied pressure/decompression rate region. Exsolution (bubble formation) may overpressure criterion accounts for the pressure (Kueppers et al., 2006a). occur at any pressure below the saturation difference between a growing bubble and the level; exsolution pressure depends on vesicu- surrounding melt (Melnik, 2000). The strain rate VOLCANIC ERUPTIONS lation kinetics. Fragmentation occurs when criterion comes from an observed threshold in the exsolved volatile phase reaches one of the deformation properties (from ductile to brittle) Eruption styles, and associated volcanic criteria discussed in the text. at high strain rates (Dingwell and Webb, 1989). landforms, were introduced descriptively in the These criteria are not mutually exclusive, and previous sections. In this section, we examine all require ascending magma to expand until the ways in which conditions of magma storage and Ductile fragmentation results from instabili- point of fragmentation. This, in turn, requires transport combine to generate some of the ob- ties in the accelerating liquid phase (e.g., Mader that the volume of gas in the component bub- served range in eruptive activity. For sim plicity, et al., 1994; Mangan and Cashman, 1996). bles increases by decompression (expansion) we separate discussions of explosive and effu- Evidence for fragmentation in the fl uid, rather and volatile exsolution faster than it escapes by sive eruptions, and then provide an overview of than solid, state comes from the fl uidal shape permeable fl ow through pathways of intercon- “transitional” eruptions, which show both ex- of mafi c volcanic bombs and commonly asso- nected bubbles (Klug and Cashman, 1996). The plosive and effusive behavior. ciated pyroclasts such as Pele’s tears (droplets) small grain size of most silicic tephra deposits, and Pele’s hair (fi nely elongated glass strands). and the uniformity of accompanying pumice Explosive Eruptions Resultant clasts in Hawaiian eruptions are textures (bubble size and number density; Fig. large—tens of centimeters—and substantially 13) suggest that fragmentation in silicic explo- Key observable parameters of witnessed ex- larger than constituent bubbles (Fig. 13). This sive eruptions is controlled primarily by bubble- plosive eruptions are the plume height (used to suggests that bubbles accelerate the magma bubble interactions (Rust and Cashman, 2011). infer eruption intensity, or mass eruption rate), (through expansion) but do not exert a direct Fragmentation can also occur in highly the duration of eruptive activity, and the fi nal control on the fragmentation process (Rust and viscous magma because of unloading by a volume, areal distribution, internal structure, Cashman, 2011). downward-propagating decompression wave and grain-size characteristics of pyroclastic Silicic melts, in contrast, are apparently frag- (Alidibirov and Dingwell, 1996; Fowler et al., deposits. Here, we briefl y review advances in mented by brittle processes. Fragmentation may 2010). Under these conditions, fragmentation understanding the dynamics of volcanic plumes occur when expanding magma exceeds a criti- may occur by: (1) propagation of an unloading and their relationship to the pyroclastic deposits cal vesicularity, when volatile phases contained elastic wave, (2) layer-by-layer bubble bursting that they produce. within bubbles attain a critical overpressure, in response to a pressure difference between the and/or when the expanding melt exceeds a criti- (pressurized) bubbles and the decompression Volcanic Plumes cal strain rate. The vesicularity criterion for si- wave, and (3) rapid gas fl ow through permea- licic Plinian eruptions has variously been placed ble networks (Alidibirov and Dingwell, 2000). Volcanic plumes form when fragmented at 60% (Kaminski and Jaupart, 1997), 64% In many situations, these mechanisms may act magma and associated gases are ejected into the (Gardner et al., 1996), and 75%–83% (Sparks, in concert. Experimental investigations of this atmosphere. For ascending hydrous magmas, 1978; Houghton and Wilson, 1989) based on process describe a minimum pressure differen- the pressure at the fragmentation level may be the observed range of vesicularity in preserved tial (fragmentation threshold) that varies with several MPa. At these pressures, the melt retains

676 Geological Society of America Bulletin, May/June 2013 How volcanoes work a substantial amount of water that may be re- sphere, and the ﬂ ow runs out of kinetic energy leased in the plume, or within density currents. and collapses to feed a pyroclastic density cur- Fragmentation, in turn, decreases the bulk vis- rent. More recent models relax the assumption HIGH Re cosity of mixture by up to 14 orders of magni- of homogeneity and allow larger particles to PARTICLES tude; the change in both bulk viscosity and bulk separate from the plume. These models show density causes the gas-particle mixture to ac- that fallout of dense particles as pyroclastic den- celerate to very high speeds (typically hundreds sity currents can increase the rise velocity of the TRANSITIONAL Re of meters per second) and to discharge into the convective plume by decreasing its bulk density PARTICLES atmosphere as a momentum-dominated jet. The (e.g., Clarke et al., 2002). Importantly, these

exit conditions of the fl ow can be divided into models also explain the common observation of Log Thickness (m) three cases that are controlled by the fl ow veloc- simultaneous production of buoyant plumes and ity and vent shape: (1) The fl ow is able to adjust pyroclastic density currents. LOW Re the atmospheric pressure at the exit; (2) the vent Volcanic plumes eventually reach a level of PARTICLES is suffi ciently narrow that the mixture exits at neutral buoyancy (HB) high in the atmosphere, above atmosphere pressure (choked fl ow) and where they spread laterally (the umbrella re- √Area (km) adjusts to ambient pressure in the atmosphere; gion). The tendency of rising plumes to over- (3) the mixture reaches supersonic speeds shoot H produces maximum plume heights H Figure 14. Schematic of deposit thickness B T √ through a diverging vent. The topic of such ~1.4H . The thermal energy fl ux determines the as a function of area covered. Linear seg- B √ fl ows is a complex area of geological fl uid me- fi nal plume height and is proportional to the in- ments on a log thickness versus area plot chanics that is far from completely understood tensity of the eruption (i.e., the fl ux of magma indicate exponential thinning; individual (for more details, see Sparks et al., 1997; Ogden through the vent in kg/s; Sparks et al., 1997). segments refl ect transitions in particle be- et al., 2008; Bercovici and Michaut, 2010). During Plinian eruptions, mass fl uxes of 105 to havior as a function of Reynolds number Once the mixture emerges as a jet into the over 109 kg/s feed eruption columns that rise to (see text). atmosphere, interaction with the air generates ≤55 km. As a result, lateral fl ow within umbrella high eruption plumes and/or pyroclastic den- regions combines with high-level winds to dis- sity currents. The former produce tephra fallout, tribute tephra-fall deposits over hundreds to mil- U is the fall velocity, dp is the particle diameter, and the latter form various kinds of pyroclastic lions of square kilometers. In contrast, weaker and µa is the viscosity of the air. The Re, in turn, surges and fl ows (and associated ash fall), which explosive eruptions generate small to moderate controls the drag coeffi cient, which controls are the most destructive and hazardous kinds of volcanic plumes with heights that are infl uenced the fall velocity. In subaerial fall deposits, par- volcanic phenomena. In both cases, the funda- strongly by wind velocity, particularly because ticle size variations exert the primary control on mental process of interaction is turbulent air of changes in conditions of air entrainment in particle Re. For this reason, the steep proximal entrainment into the high-speed erupting mix- bent-over plumes (Bursik, 2001). As a conse- segment shown in Figure 14 can be attributed to ture; this has two major consequences. First, quence, the mass fl ux required for weak plumes fall of large particles (high Re) from the outer entrainment of air decelerates the ascending to reach a specifi c height increases markedly as margin of the rising plume, the middle segment mixture by momentum transfer (the entrained wind speed increases. to fall of intermediate-size particles (transitional air has to be accelerated). Second, the entrained Re) from the umbrella region, and the distal seg- air is heated, and the mixture density reduces Pyroclastic Fall Deposits ment to deposition of fi ne particles in a low-Re as the plume rises. As erupting mixtures are regime (e.g., Bonadonna et al., 1998; Alfano almost always denser than air, they typically The characteristics of pyroclastic fall depos- et al., 2011). The volume of the distal segment have enough initial kinetic energy to rise only its can be related to the nature of the eruption is the most diffi cult to quantify accurately, be- hundreds of meters to a few kilometers into the plumes that produced them. For this reason, cause fi ne ash layers are poorly preserved and atmosphere. Thus, formation of the towering pyroclastic fall deposits are commonly used to distributed over vast areas of land and ocean. convecting eruption columns commonly seen assign both magnitude and intensity to prehis- In submarine environments, density becomes a in Plinian eruptions requires air entrainment and toric eruption deposits, information that is criti- critical parameter in determining conditions of heating of the air by the volcanic particles. This cal for volcanic hazard assessment. particle deposition (Cashman and Fiske, 1991); process generates potential energy by convert- Deposit magnitude is measured by changes in the density difference between pumice and sea- ing thermal energy in the magma to mechanical deposit thickness (or, ideally, mass) as a func- water will vary greatly depending on whether energy through buoyancy of the mixture. Typi- tion of distance from the vent. Tephra depos- the pore spaces within pumice are fi lled with air, cally, thermal energy is more than an order of its generally thin exponentially away from the steam, or water (Allen et al., 2008). magnitude greater than the kinetic energy. source vent and can be used to estimate the total Eruption intensity (mass eruption rate) can be Early model treatment of the erupting mixture volume (or mass, if corrected for deposit den- inferred from derived relationships among grain of gas and entrained material as a homogeneous sity) of a fall deposit (Pyle, 1989). Log thick- size, grain density, column height, and deposi- “pseudogas” led to the development of two end- ness versus distance (measured as √area) plots tional characteristics (e.g., Carey and Sparks, member scenarios (Woods, 1995): (1) Heating for Plinian deposits are thus linear, although 1986; Ernst et al., 1996; Burden et al., 2011). of entrained air makes the mixture less dense they commonly show three or four different Although simple in principle, large uncertain- than the surrounding atmosphere, and the jet linear segments because of variations in fall ties are introduced during the collection of fi eld transforms into a buoyant plume to form a high behavior (e.g., Fierstein and Nathenson, 1992; data (e.g., Biass and Bonadonna, 2011). Per- eruption column in the atmosphere; or (2) air en- Fig. 14). The fall behavior (terminal velocity) is haps more important is characterization of the trainment is not suffi cient to reduce the density controlled by the Reynolds number (Re), where total grain-size distribution (TGSD), which is ρ ρ of the erupting mixture below that of the atmo- Re = PUdp/µa; here p is the particle density, critical input for ash dispersion models (Mastin

Geological Society of America Bulletin, May/June 2013 677 Cashman and Sparks et al., 2009a). TGSDs are notoriously diffi cult preserves evidence of multiple discrete fl ows Another important characteristic of many to measure accurately because (1) tephra depos- separated by time breaks of minutes to hours pumiceous pyroclastic fl ows is their extreme its are widespread, (2) the grain size at a single (Fierstein and Wilson, 2005). In the example of mobility, even on very low slopes (e.g., Druitt, site often varies up section, thus requiring mul- the Bishop Tuff, California, a coeval fall deposit 1998). Recognition of this mobility has led to tiple measurements at individual locations, and has been used to infer an emplacement time of experimental studies of fl ow behavior as in- (3) fi ne-grained distal deposits are often poorly ~90 h for the entire fl ow-fall sequence (Wilson fl uenced by fl uidization, gas retention, pore- preserved and/or limited in accessibility (e.g., and Hildreth, 1997). pressure generation, and depositional processes deposited within the ocean). For this reason, Field characterization of pyroclastic depos- (e.g., Dellino et al., 2010; Roche et al., 2010; there are relatively few tephra deposits for which its is often complicated by postemplacement Girolami et al., 2010). Fluidization occurs when the TGSD is well constrained. A compilation modifi cation of primary ignimbrite textures, the weight of a particle is balanced by the ver- of existing data shows sharp differences among particularly those related to welding. Welding tical drag force exerted by a fl owing gas, and deposits of different types, with Plinian/subplin- refers to the densifi cation (porosity reduction) is therefore determined by the settling velocity ian deposits typically fi ne grained and poorly of pyroclastic fl ow deposits by a combination of individual particles (Fig. 15A). A particulate sorted, Vulcanian/violent Strombolian eruptions of sintering, compaction, and fl attening of con- bed expanded by fl uidization will collapse when medium grained and often well sorted (particu- stituent material. These physical changes refl ect the gas supply is reduced, causing the particles larly the mafi c end members), and Strombo- the weight of overlying material, the viscosity to be deposited (Fig. 15B). When viewed from lian/Hawaiian deposits coarse grained and well and porosity of the deposit (controlled by erup- another perspective, the gas retention (and resul- sorted (Fig. 13). Data of this type are critical for tion conditions, composition, and tempera- tant high pore pressures) within a fl uidized fl ow developing nominal source input parameters ture), and the time available for deformation (a required for pyroclastic density current mobility for plume models (e.g., Mastin et al., 2009a). function of the rates of cooling and gas loss). will be controlled by time scales for both dif- For the end-member case of no volatile resorp- fusive outgassing and hindered settling of con- Pyroclastic Density Currents tion, the degree of compaction is limited by the stituent particles (e.g., Druitt et al., 2007). Also permeability of the deposit (e.g., Riehle et al., important is the trajectory of particles and gas Pyroclastic density currents are among the 1995). However, in thick deposits that accumu- within the moving fl ow (e.g., Giordano, 1998). most hazardous of all volcanic phenomena, and late rapidly, water vapor can dissolve into glass From a broader perspective, theoretical and yet they are also one of the least understood. and facilitate pore-space collapse if pore pres- experimental studies of multiphase flows Pyro clastic density currents are hot gravity- sures are suffi ciently elevated and permeability show that mixtures of high-temperature solids, driven currents that travel at high velocities and is suffi ciently low (Sparks et al., 1999). Welding liquids, and gases can simultaneously have inundate (and bury) large areas. They can form characteristics thus place important constraints properties of gas (e.g., compressibility), of by lava dome collapse, by column collapse dur- on fl ow emplacement conditions. solids (through small-scale interactions between ing Plinian/subplinian eruptions, or accompa- From a process perspective, pyroclastic den- particles), and of liquids. Moreover, individual nying caldera collapse during large-magnitude sity current deposits can be viewed as either in- particles can behave like rigid solids or ductile explosive eruptions. The high velocities, high crementally deposited from the density current liquids, depending on time scales of deforma- temperatures, and complex nature of these fl ows (e.g., Branney and Kokelaar, 2003) or deposited tion. For these reasons, most models use end- make it impossible to measure either their ma- rapidly when the fl ow loses energy (e.g., Wilson, member descriptions and treat pyroclastic terial properties or their dynamics directly. For 1985). It is likely that these differences of view density currents as either a turbulent suspension this reason, pyroclastic density current studies refl ect the real complexities of these high-tem- (dilute) or a granular fl ow (dense). Granular combine fi eld observations of deposits (ignim- perature multiphase fl ows. One way to combine fl ow dynamics, including fl uidization by escap- brites) with laboratory experiments and numeri- these two perspectives is to view pyroclastic ing gases, provide a framework for modeling cal models of fl ow dynamics. density currents as having two components, the concentrated parts of pyroclastic density Detailed studies of individual ignimbrites with overlying dilute turbulent ash-cloud surges currents (e.g., Titan2D; Patra et al., 2005). Other show that neither the grain size nor the com- and concentrated dense basal fl ows (e.g., Druitt, important factors are interactions with topog- positional variation of the ignimbrite at any 1998). Pyroclastic density currents can then be raphy that cause the dense basal portion of the given location can be simply related to the na- described as a continuum between these two fl ow to decouple from the dilute turbulent cloud ture of either the eruptive mixture or the parent end members. This conceptual model allows the (e.g., Andrews and Manga, 2011; Esposti On- fl ow (Wilson, 1985; Branney and Kokelaar, mass distributed between the two components to garo et al., 2011). Modeling the dense basal 1992, 2003). Instead, ignimbrites show distinct vary in space and time, or even to decouple from current is critical for predicting maximum run- facies that result from transport and deposition one another. Such fl ow transformations (be- out distances of pyroclastic density currents processes. An important factor from a hazards tween dilute and dense fl ows) are often caused (Doyle et al., 2008). Limitations of existing point of view is using the physical attributes of by topographic changes and have major hazards models are highlighted by fi eld evidence for a ignimbrites to constrain eruption time scales. implications (e.g., Giordano, 1998; Druitt et al., wide range of emplacement temperatures (e.g., For example, in the low-aspect-ratio Taupo 2002a). A third component of pyroclastic den- McClelland et al., 2004; Gurioli et al., 2005; ignimbrite, lateral variations in grain-size dis- sity currents is the overlying buoyant ash plume Lesti et al., 2011), which point to the need to tribution are similar to vertical variations in a that develops in parts of the current that become incorporate effects of temperature into existing fl uidized bed, which suggests that fl uidization less dense than the overlying atmosphere (e.g., models. Moreover, although separate models of was critical to fl ow emplacement and, by exten- Calder et al., 1999). Co-fl ow, or coignimbrite, dilute and dense fl ow components are useful for sion, that emplacement was rapid and occurred ash plumes can be generated either continuously understanding individual end members, future as a discrete event (Wilson, 1985). In contrast, during fl ow or abruptly, if the pyroclastic den- models must strive to couple these two different a high-aspect-ratio valley-fi lling ignimbrite gen- sity current is initially well mixed but suddenly regions to fully describe pyroclastic density cur- erated by the 1912 eruption of Katmai, Alaska, loses mass by deposition and becomes buoyant. rent behavior (e.g., Neri et al., 2003).

678 Geological Society of America Bulletin, May/June 2013 How volcanoes work

A Expansion fl ows. The rates and mechanisms of fl ow ad- Total bed height vance are therefore controlled primarily by de- (dense phase + bubbles) velopment of a solid crust on the fl ow surface H mb (Griffi ths, 2000), as well as by cooling-induced Dense phase height (H be ) crystallization of fl ow interiors (Cashman et al., 1999). In contrast, lava fl ows produced by more water-rich magmas, such as those of Mt. Etna, BUBBLING REGIME Italy, experience syneruptive crystallization because of volatile loss during magma ascent, and Bed Height may therefore be highly crystalline on eruption. H mp These fl ows have higher viscosities, and are UNIFORM REGIME shorter and advance more slowly, than Hawaiian lava fl ows (e.g., Kilburn, 2004). Common to U U mp mb Gas velocity both regions, however, are pahoehoe and ‘a‘ā morphologies, surface textures that refl ect both B Collapse rheological and deformation rate thresholds (1) (2) (3) (Figs. 16A–16C). These morphological differ- (1) BUBBLE ences may also be viewed from the perspective U EVACUATION sett of simple and compound fl ow forms.

Bed Height Simple lava fl ows may have lengths that are H be (2) HINDERED SETTLING (slope = U ) limited either by erupted volume or by cooling. settled When simple fl ows are cooling-limited, they particles are generally assumed to have lengths that are proportional to the extrusion rate (Walker et al., 1973; Harris et al., 2007), although this correla- sett (3) PACKED BED tion is not always robust for Hawaiian lava fl ows H sett (e.g., Riker et al., 2009). Also important for haz- Time ard assessment is the recognition that higher- effusion-rate fl ows advance more rapidly than Figure 15. Behavior of fl uidized beds. (A) Schematic diagram of the lower-effusion-rate fl ows (Rowland and Walker,

expansion regime, showing the onset of fl uidization at Ump (gas veloc- 1990; Kauahikaua et al., 2003), and that fl ow ad- ity that creates maximum pressure drop tolerated by the particle bed) vance rates diminish with the distance that a fl ow

and the onset of the bubbling regime at gas velocity Ube. When Ump has traveled. Together, these constraints indicate

< U < Ube, the bed expands uniformly (although the distribution of that lava fl ow hazards are determined by both ini- gas in this regime depends on the particle size, shape, and density dis- tial rates of effusion and proximity to vent regions tribution). (B) Schematic diagram of the collapse regime, which oc- (e.g., Kauahikaua et al., 2002; Soule et al., 2004). curs when the gas fl ow is reduced. Initially, gas is lost as bubbles rise Another characteristic of simple lava fl ows is

through the bed. When H = Hbe, all macroscopic bubbles have escaped, that they form channels by construction of lat-

and the bed slowly densifi es as particles settle. Umb is the gas velocity eral levees. Crystal-rich Etna lavas have a yield at which bubbling starts. Figure is modifi ed from Druitt et al. (2007). strength that promotes channel (and levee) development by inhibiting lateral spreading (Hulme, 1974). In contrast, fl uid Hawaiian fl ows cool Effusive Eruptions Mafi c Lava Flows rapidly, so that levees develop because spreading is inhibited by solidifi cation at the fl ow margins Lava fl ows form when magma degassing is The two primary laboratories for studies (Kerr et al., 2006). An understanding of controls suffi ciently fast relative to the rate of ascent that of mafi c lava fl ows over the past two decades on channel geometries is required for develop- magma reaches the surface without fragment- have been Hawaii, United States (e.g., Heliker ing fully predictive models of lava fl ow advance ing. This may occur by slow ascent of H2O-rich et al., 2003), and Etna, Italy (e.g., Bonaccorso (e.g., Harris and Rowland, 2001). Challenges to magma accompanied by gas loss, or more rapid et al., 2004). Both have had frequent (or in the modeling include not only the common presence

rise of H2O-poor magma. Alternatively, clasto- case of Hawaii, continuous) eruptive activ- of multiple parallel channels (e.g., Favalli et al., genic lava fl ows may form from re-fusion of ity over this time period; as a result, both have 2010; James et al., 2010) but also the distribu- fragmented material. Individual lava fl ows may contributed extensive observational data sets of tary nature of many channelized fl ows, which range in volume from a few cubic meters to hun- lava fl ow behavior. Moreover, as Hawaii erupts split because of topographic barriers and chan- dreds (or even a few thousand) cubic kilometers H2O-poor magma while Etna erupts hydrous nel overfl ows produced by temporary increases in large fl ood basalt eruptions. The range of magma, comparisons of these two systems al- in lava supply or channel blockages. fl ow emplacement conditions is refl ected in the low assessment of the role of water (particularly Compound lava fl ows may consist of tens variability of fl ow morphology, length, thick- degassing-induced crystallization) in lava fl ow to thousands of individual lava lobes and are ness, structures, and surface textures. As with emplacement. most common in large and long-lived tube-fed explosive eruptions, we consider separately the In general, Hawaiian lava erupts at near- pahoehoe fl ow fi elds, such as the Pu’u ‘Ō’ō- behavior of mafi c and silicic lavas. liquidus temperatures and cools rapidly as it Kūpaianaha fl ow fi eld that has developed in

Geological Society of America Bulletin, May/June 2013 679 Cashman and Sparks

Figure 16. Hawaiian lava ﬂ ows. 3 TRANSITI (A) Pahoehoe lobe showing A C rapid cooling, skin formation,

ON THRESHOLD and surface deformation dur- 2 ing emplacement. (B) ‘A‘ā fl ow 100 µm with broken crust and exposed ’A’ā lava core. (C) Strain rate–ap- 1 parent viscosity relationships Shear rate (1/s) ZONE and estimated transitional Pāhoehoe threshold zone (TTZ) separat- ing pahoehoe from ‘a‘ā (modi- 0 fied from Hon et al., 2003; 23 4 Log apparent viscosity (Pa s) Soule and Cashman, 2005). ) Inset photographs are back- B Ψ 1 D MOBILE CRUST scattered electron images of (channelized ’a’ā) quenched pahoehoe and ‘a‘ā lavas; dark-gray phase is plagioclase, medium-gray phase is pyroxene, light-gray phase 0 is glass; scale bar is the same for both images. (D) Parameter- ization from analogue experi- Ψ TUBES ments showing threshold in (tube-fed pāhoehoe) (advection time scale/cooling –1 time scale)–Ra (Rayleigh num- Advection rate / Cooling ( Log 4 56 789 ber measures thermal convec- Log Ra tion strength) space. Open squares represent experiments in “mobile crust” (open channel) regime; fi lled circles show experiments in tube regime; line is boundary between regimes (modifi ed from Griffi ths et al., 2003).

Hawaii between 1986 and the present (Heliker A physics-based description of cooling-lim- et al., 2004). Satellite-based radar images have et al., 2003). Lava tubes form where a solidifi ed ited behavior is that the fl ow has reached a criti- the advantages of both seeing through cloud roof is created and maintained over a section of cal Peclet number, which is the ratio of the fl ow cover and having higher resolution than satellite- the fl ow (e.g., Peterson et al., 1994; Cashman rate to the product of the thermal diffusivity and based thermal imaging techniques. Radar cor- et al., 2006). Because solidifi ed lava has low fl ow distance (e.g., Pinkerton and Wilson, 1994; relation imaging, in particular, provides image thermal conductivity, lava tubes are well insu- Kerr and Lyman, 2007). An extension of this resolution suffi cient for monitoring individual lated and facilitate lava transport over large dis- concept shows that fl ow surface morphology is lava fl ows (e.g., Zebker et al., 1996; Dietterich tances with little cooling (e.g., Helz et al., 1995, governed by the balance between fl ow cooling et al., 2012), as well as postemplacement fl ow Ψ 2003; Ho and Cashman, 1997). Two important and fl ow advection (measured as = Uots/Ho, volumes (e.g., Stevens et al., 1997; Lu et al., phenomena in tube-fed compound lavas are fl ow where Uo is fl ow velocity, ts is cooling time, and 2003) and cooling-induced subsidence (Stevens infl ation and lava tube drainage. On low slopes, Ho is fl ow thickness), and the strength of internal et al., 2001). High-resolution thermal imaging lava tubes form within spreading sheet fl ows. The convection within the fl ow (as measured by the data can be obtained using airborne (e.g., Real- tubes are typically fi lled with lava, such that con- Rayleigh number Ra; Fig. 15D). When cooling muto et al., 1992) and hand-held (e.g., Harris tinued fl ow through the tube is accompanied by rates are large relative to fl ow advance, an insu- et al., 2005; Ball and Pinkerton, 2006; Spampi- cooling-induced lava accretion to the tube roof. lating surface crust forms to produce lava tubes nato et al., 2011) cameras. Similarly airborne Under these conditions, maintenance of a con- that feed pāhoehoe fl ows; when fl ow advance (ALSM) and terrestrial (TLS) laser scanning stant lava fl ux through the tubes requires that the is rapid, the insulating crust is continually dis- and ground-based radar provide high-resolution fl ow infl ate (Hon et al., 1994). Thus, Hawaiian rupted, the interior lava cools rapidly, and ‘a‘ā digital topographic data. In all cases, spatial lava fl ows with initial heights of only centime- fl ows are formed. resolution is improved at the expense of the ters commonly infl ate to several meters during The past few decades have seen an explo- aerial (and often temporal) coverage of satellite- prolonged emplacement. On steeper slopes, lava sion of new tools applied to lava fl ow studies, based systems. These data are revolutionizing tubes are not completely fi lled, and persistent including global positioning system (GPS), digi- quantitative analysis of lava fl ows, as they allow fl ow in large tubes can lower the tube fl oor by tal topographic data, and satellite-based remote detailed imaging of the thermal and morpho logi- thermal and mechanical erosion (Kauahikaua sensing. Appropriate application of these tools cal evolution of lava fl ows that can be related et al., 1998; Kerr, 2001). When the lava supply requires balancing the spatial and temporal reso- to the dynamics of emplacement (e.g., Harris rate diminishes, lava tubes may drain, sometimes lution with the areal coverage. Satellite-based et al., 2007). Particularly exciting is the advent creating a series of collapse pits. These features thermal images generally have low spatial (1–4 of multi temporal imaging of active fl ows (e.g., are common in older volcanic landscapes, and km/pixel) but high temporal resolution, and are Favalli et al., 2010; James et al., 2010), which have also been recognized on both the Moon therefore used for monitoring entire fl ow fi elds provides detailed information on the develop- and Mars (e.g., Garry and Bleacher, 2011). (for reviews, see Oppenheimer, 1998; Wright ment of individual lava fl ow channels and lobes.

680 Geological Society of America Bulletin, May/June 2013 How volcanoes work

Together, these new measurement capabilities tions fed by magma supply rates intermediate 1.0 SUBPLINIAN can be used to test proposed models of channel between those of the explosive and effusive development, lava tube formation, rates of fl ow counterparts. Mafi c transitional eruptions often 0.8 advance, and fl ow conditions within lava chan- show simultaneous Strombolian/violent Strom- 0.6 nels; they also provide new ways to assess the bolian explosions from a central scoria cone and VIOLENT 0.4 hazard and risk posed by lava fl ow inundation. lava effusion from the cone base. A classic ex- STROMBOLIAN ample of this type of activity is the 1943–1952 0.2

Viscous Flows and Domes eruption of Paricutin volcano, Mexico (Luhr mass Total Mass of tephra / EFFUSIVE and Simkin, 1993). Silicic transitional erup- 0 1 3 5 7 9 Well-studied examples of viscous lava fl ows tions include alternation between lava dome/ Log mass eruption rate (kg/s) and domes are provided by recent activity at plug formation and Vulcanian to subplinian ex- Mount St. Helens, United States (1980–1986 plosions, as seen at Mount St. Helens in 1980 Figure 17. Value of tephra production (mass and 2004–2008; Swanson and Holcomb, 1990; (e.g., Cashman and McConnell, 2005) and at tephra/ mass[tephra + scoria cone + lava]) as a function Sherrod et al., 2008), Unzen, Japan (1991– Soufriere Hills, Montserrat in 1996 (e.g., Voight of mass eruption rate (in kg/s). Filled circles 1995; e.g., Nakada and Motomura, 1999), and et al., 1999). Although these eruptive patterns do show annual data from the 1943–1952 Soufriere Hills, Montserrat (1995–present; not fi t neatly into simple classifi cation schemes, eruption of Parícutin, Mexico; open circles e.g., Voight and Sparks, 2010). In all of these understanding transitions in eruptive behavior is represent other mafi c eruptions. Figure is examples, degassing and crystallization during critical for improved hazard assessment during modifi ed from Pioli et al. (2009). magma ascent cause very large increases in vis- volcanic crises. cosity, and lava dome morphology is controlled Eruptions of low-viscosity mafi c magma vary by magma ascent rate through the kinetics of the in explosivity with changes in rates of magma constraints on the rates of gas loss, densifi cation, phase changes. ascent. When the magma ascent rate is negli- and pressure buildup (e.g., Druitt et al., 2002b). When magma ascent is rapid, volatiles are gible, rising gas bubbles can coalesce and reach Vulcanian activity can trigger subplinian erup- retained within the melt, and crystallization is the surface as large bubble bursts, as commonly tions when the initial downward-propagating limited. In this case, the relatively fl uid erupt- seen in lava lakes and open vent volcanoes decompression wave produced by the Vulcanian ing magma creates either pancake-like domes such as Villarica, Chile, Sanguy, Ecuador, and event triggers degassing and eruption of gas-rich (e.g., Watts et al., 2002) or obsidian fl ows. the eponymous Stromboli volcano, Italy. When magma within the conduit. Both Vulcanian and Scaling analysis suggests that eruption rates of magma ascent rates are higher, lava fl ows may subplinian eruptions are often associated with ~10–100 m3/s may be required to produce these emerge from lateral vents (e.g., Ripepe et al., effusion of lava fl ows or domes. fl ow morphologies (e.g., Lyman et al., 2004). 2005) or initiate simultaneous effusive and ex- Transitional activity at intermediate-compo- These rates are consistent with strain rates ob- plosive activity that characterizes violent Strom- sition volcanoes may alternate between Strom- tained from microlite orientations (Castro et al., bolian eruptions (Pioli et al., 2008). At even bolian and Vulcanian explosions depending on 2002) and with recent observations of extrusion higher rates of magma ascent, eruptions are the rate of magma supply, which controls the ex- rates of 20–100 m3/s during the fi rst few months subplinian. This progression refl ects a decrease tent of crystallization and, as a result, the com- of rhyolite lava extrusion during the 2008–2009 in the effi ciency of synascent gas segregation position of the matrix melt (Wright et al., 2012). eruption of Chaiten volcano, Chile (Carn et al., (and resulting increase in eruption explosivity Examples include ongoing activity at Tungura- 2009). The dense and degassed nature of obsid- and tephra production) with increased rates of hua volcano, Ecuador, and Llaima, Chile (http:// ian further requires effi cient gas loss via both magma ascent (Fig. 17). www.volcano.si.edu). This alternation can be gas fl ow through a permeable foam (Eichel- Intermediate/silicic magmas also have erup- explained by variations in rates of magma as- berger et al., 1986) and along permeable and tive styles that refl ect the rate of magma supply cent, and resulting changes in crystallinity, melt fractured conduit walls (e.g., Tuffen et al., 2003; to the vent. In this case, eruptive style changes composition, and mode of gas escape (Fig. Rust et al., 2004; Cabrera et al., 2011). from Vulcanian to subplinian and fi nally to sus- 11A). Melt composition appears particularly When magma ascent is very slow, degassing tained Plinian explosive eruptions as the magma important, as illustrated by a plot of sample and crystallization combine to produce mag- supply rate to the vent increases. The change in crystallinity (phenocryst, microphenocryst, and mas with high viscosity and non-Newtonian eruptive style probably refl ects the effi ciency of groundmass) and matrix glass composition (as rheologies. In the extreme, the lava solidifi es both degassing and degassing-induced crystal- wt% SiO2; Fig. 18). This plot shows that the completely and extrudes as a rigid spine with lization relative to the velocity of magma ascent products of volcanoes that exhibit transitional marginal fault zones (e.g., Cashman et al., 2008; (Cashman, 2004; Mason et al., 2006). Of this activity are typically crystal-rich, and that the Pallister et al., 2008; Fig. 3D). Thus, a wide spectrum, Vulcanian activity requires the most eruptive style (violent Strombolian, Vulcanian, spectrum of lava morphologies, from pancake- effi cient gas loss from, and densifi cation of, or both) is strongly dependent on the residual shaped domes to shear lobes and spines, can be magma residing within shallow conduits (e.g., melt composition. explained simply by variations in effusion rate Fig. 11B). The nature of the resulting plug, and and resulting changes in bulk rheology (e.g., the characteristics of conduit-fi lling magma be- Eruptions Involving Water Nakada and Motomura, 1999; Watts et al., 2002). tween eruptions, can be determined by combining analysis of ejected breadcrust bombs (e.g., All of the eruptive behaviors reviewed herein Transitional Eruptions Wright et al., 2007) with models that link bomb relate to purely magmatic activity, that is, eruptive characteristics to conduit pressure (e.g., Bur- behavior that is controlled entirely by the physi- Eruptions may be considered transitional gisser et al., 2011). Information on the repose cal properties and driving forces of the magma when they include both explosive and effusive interval between events and the volume erupted itself. However, rising magma may also encoun- activity; transitional activity characterizes erup- during each explosion can place important time ter either groundwater or surface water/snow/ice

Geological Society of America Bulletin, May/June 2013 681 Cashman and Sparks

Figure 18. Plot of pyroclast 90 STROMBOLIAN VULCANIAN tensive fragmentation requires intimate mixing crystallinity (phenocrysts, micro- between magma and water. This is generally phenocrysts, and microlites) GAL most effi cient for low-viscosity mafi c magmas COL as a function of SiO2 (wt%) (e.g., Zimanowski et al., 2003), although recent 60 TUN PCH in the groundmass glass for SH MRT experiments suggest that shear-induced changes products of transitional erup- ET MSH PAR in melt deformation from ductile to brittle may tions; eruptive style is labeled CN facilitate water interaction in more viscous as Strombolian (includes vio- 30 melts (Austin-Erickson et al., 2011). Both the lent Strombolian) and Vul- Crystallinity (%) VS infl uence of water on changes in eruption style canian. All eruptive products and the ability of external water to increase the are highly crystalline; eruption effi ciency of magmatic fragmentation are im- style appears to refl ect matrix 040 50 60 70 80 portant for evaluating volcanic hazards in re- glass composition (viscosity). SiO2 in matrix glass (wt%) gions where there is potential for magma-water Volcanoes that erupt magma interaction, particularly regions of distributed with bulk compositions of basaltic andesite to andesite can have different eruptive styles volcanism (such as cinder cone fi elds). An area depending on the amount of groundmass crystallization (and resulting evolution of matrix of future research is in developing ways to map glass). ET—Etna, Italy; VS—Vesuvius, Italy; CN—Cerro Negro, Nicaragua; SH—Shishaldin, both the spatial extent of groundwater systems, United States; PAR—Paricutin, Mexico; TUN—Tungurahua, Ecuador; GAL—Galeras, and the effective permeability of host rocks in Colombia; COL—Colima, Mexico; PCH—Pichincha, Ecuador; MRT—Soufriere Hills, active volcanic regions. Montserrat; MSH—Mount St. Helens, United States. Data are from Santacroce et al. (1993), Calvache and Williams (1997), Roggensack et al. (1997), Cashman and Blundy (2000), Luhr CONCLUDING REMARKS (2001), Stelling et al. (2002), Mora et al. (2002), Harford et al. (2003), Tadduecci et al. (2004), and Wright et al. (2012). The overview provided here highlights the physical processes responsible for magma ascent, arrest in the upper crust, migration to- during ascent and eruption. The past few decades on subaqueous eruption-fed density currents ward Earth’s surface, eruption, and emplace- have seen numerous advances in fi eld, experi- (White, 2000), historic submarine pumice erup- ment as pyroclastic fall and fl ow deposits, or as mental, and theoretical perspectives on magma- tions (Kano, 2003), and explosive sub marine lava fl ows and domes. Critical factors to most water interactions, as reviewed in Head and eruptions (White et al., 2003) discuss eruption of these processes is the behavior of volatile Wilson (2003) and White et al. (2003). and deposition of pumice on the seafl oor. A phases as they exsolve and escape from the In submarine environments, the style of critical factor for understanding pumice deposi- transporting magma, the response of the melt magma-water interaction depends primarily on tion is defi ning conditions under which pumice phase in terms of phase stability and crystalli- the height (pressure) of the overlying water col- transforms from being less dense to being more zation, and the resulting changes in rheology. umn and its effect on vapor formation and ex- dense than the surrounding seawater (e.g., Cash- These complex interactions speak to the need pansion (e.g., Kokelaar, 1986; Head and Wilson, man and Fiske, 1991; Allen et al., 2008). for continued advances in our understanding of 2003). Although the overlying water pressure is Magma-water interactions in subaerial envi- both physical and chemical aspects of the entire unlikely to affect fragmentation of ascending ronments include interaction of rising magma phase space that encompasses mixtures of gases, hydrous magmas (which should reach fragmen- with groundwater aquifers, hydrothermal sys- liquids, and particles. Additionally, because an tation conditions well within the conduit), the tems, or surface water (including ice and snow). important goal of volcanological research is water column will control the extent to which Interaction of magma with external water typi- improved assessment of the hazards posed by eruption plumes are suppressed by the weight cally produces abundant fi ne ash, with the small volcanic eruptions and the associated risks to of the overlying water column. Interaction with grain size refl ecting the high energy provided by human populations, improved understanding of seawater can also enhance magmatic fragmenta- water expansion (e.g., Koyaguchi and Woods, the underlying processes must be translated to tion by rapid quenching. 1996; Mastin, 2007). Introduction of exter- improvements in hazard and risk assessment. Observations from many mid-ocean-ridge nal water may affect the course of magmatic For this reason, the past few decades have seen and ocean-island environments show extensive eruptions, as illustrated by the 1875 eruption a proliferation of cross-disciplinary research evidence for fragmentation driven by magmatic of Askja volcano, Iceland, where variations themes and publications related to volcanic gases, including characteristic mafi c fl uidal in groundwater availability controlled shifts in activity, including volcano impacts on health pyro clast forms such as Pele’s hair (e.g., Davis eruption style (Lupi et al., 2011). (e.g., Hansell and Oppenheimer, 2004; Horwell and Clague, 2006; Clague et al., 2009). In con- Pyroclasts produced by phreatomagmatic and Baxter, 2006), culture (e.g., Cashman and trast, hydrous mafi c magmas that have experi- eruptions often have lower vesicularities than Giordano, 2008; Grattan and Torrence, 2010), enced both degassing and crystallization during pyroclasts from magmatic eruptions, a char- religion (e.g., Gaillard and Texier, 2010), and ascent appear more susceptible to secondary acteristic that is attributed to premature clast societal resilience (e.g., Paton and Johnston, (quench fragmentation) processes (e.g., Dear- quenching because of water (e.g., Houghton and 2006), as well as modeling of ash plumes (e.g., dorff et al., 2011). Silicic submarine eruptions Wilson, 1989). The angular form of many clasts Mastin et al., 2009a) and issues of risk and un- may be highly explosive, as shown by silicic cal- also points to the importance of quench frag- certainty in volcanic hazard assessment (Sparks deras on the modern seafl oor that have produced mentation, particularly of rapidly chilled glassy et al., 2012). Although a thorough review of substantial volumes of highly vesicular pumice rinds (Mastin et al., 2009b), as does the appear- these themes is outside the scope of this re- deposits (e.g., Fiske et al., 2001; Wright et al., ance of quench cracks on particle surfaces (e.g., view, we end by placing basic volcanological 2003; Tani et al., 2008). Recent review papers Büttner et al., 1999; Dellino et al., 2012). Ex- research within the context of applied research

682 Geological Society of America Bulletin, May/June 2013 How volcanoes work to illustrate the challenges posed by the need ACKNOWLEDGMENTS Geothermal Research, v. 149, p. 85–102, doi:10.1016 for both short- and long-term forecasts of vol- /j.jvolgeores.2005.06.002. KVC would like to acknowledge support from the Bacon, C.R., 1983, Eruptive history of Mount Mazama and canic activity. AXA Research Fund; RSJS acknowledges support Crater Lake caldera, Cascade Range, U.S.A.: Journal The goal of short-term eruption forecasting is from the European Research Council. We also thank of Volcanology and Geothermal Research, v. 18, p. 57– 115, doi:10.1016/0377-0273(83)90004-5. to estimate eruption likelihood during periods of Guido Giordano and Clive Oppenheimer for helpful Bacon, C.R., Gardner, J.V., Mayer, L.A., Buktenica, M.W., obvious volcanic unrest. Challenges relate to the feedback on the fi rst draft, and Brendan Murphy for Dartnell, P., Ramsey, D.W., and Robinson, J.E., 2002, complexity of natural processes and the uncer- suggesting that we write this paper. Morphology, volcanism, and mass wasting in Crater Lake, Oregon: Geological Society of America Bulletin, tainties inherent in trying to interpret monitor- REFERENCES CITED v. 114, p. 675–692, doi:10.1130/0016-7606(2002)114 ing signals when the source of those signals lies <0675:MVAMWI>2.0.CO;2. well below Earth’s surface (and is therefore not Acocella, V., 2007, Understanding caldera structure and de- Ball, M., and Pinkerton, H., 2006, Factors affecting the velopment: An overview of analogue models compared accu racy of thermal imaging cameras in volcanology: observable). Forecasting is most accurate when to natural calderas: Earth-Science Reviews, v. 85, Journal of Geophysical Research, v. 111, B11203, doi: multiple types of monitoring data are available, p. 125–160, doi:10.1016/j.earscirev.2007.08.004. 10.1029/2005JB003829. although detailed predictions of the time, place, Acocella, V., Cifelli, F., and Funiciello, R., 2000, Analogue Barmin, A., Melnik, O., and Sparks, R.S.J., 2002, Periodic models of collapse calderas and resurgent domes: Jour- behaviour in lava dome eruptions: Earth and Plan- and type of eruptive activity are diffi cult even nal of Volcanology and Geothermal Research, v. 104, etary Science Letters, v. 199, p. 173–184, doi:10.1016 with accurate data. However, as the number p. 81–96, doi:10.1016/S0377-0273(00)00201-8. /S0012-821X(02)00557-5. Barnett, W.P., and Lorig, L., 2007, A model for stress- of active volcanoes with good monitoring net- Alfano, F., Bonadonna, C., Delmelle, P., and Costantini, L., 2011, Insights on tephra settling velocity from mor- controlled pipe growth: Journal of Volcanology and works increases, the hope is that patterns of pre- phological observations: Journal of Volcanology and Geothermal Research, v. 159, p. 108–125, doi:10.1016 eruptive activity can be recognized and applied Geothermal Research, v. 208, p. 86–98, doi:10.1016 /j.jvolgeores.2006.06.006. /j.jvolgeores.2011.09.013. Belien, I.B., Cashman, K.V., and Rempel, A.W., 2010, Gas to interpretation of future events. In this regard, Alidibirov, M., and Dingwell, D.B., 1996, Magma fragmen- accumulation in particle-rich suspensions and implica- there is a push toward developing global data- tation by rapid decompression: Nature, v. 380, p. 146– tions for bubble populations in crystal-rich magma: bases such as WOVOdata, the World Organiza- 148, doi:10.1038/380146a0. Earth and Planetary Science Letters, v. 297, p. 133– Alidibirov, M., and Dingwell, D.B., 2000, Three fragmen- 140, doi:10.1016/j.epsl.2010.06.014. tion of Volcano Observatories database (http:// tation mechanisms for highly viscous magma under Bell, A.F., and Kilburn, C.R.J., 2012, Precursors to dyke-fed www.wovodat.org/), and GVM, the Global Vol- rapid decompression: Journal of Volcanology and eruptions at basaltic volcanoes: Insights from patterns of volcano-tectonic seismicity at Kilauea volcano, cano Model (http://www.globalvolcanomodel Geothermal Research, v. 100, p. 413–421, doi:10.1016 /S0377-0273(00)00149-9. Hawaii: Bulletin of Volcanology, v. 74, doi:10.1007 .org/). These databases can be used not only for Allen, S.R., Fiske, R.S., and Cashman, K.V., 2008, Quench- /s00445-011-0519-3. pattern recognition, but also to provide input ing of steam-charged pumice: Implications for sub- Bercovici, D., and Michaut, C., 2010, Two phase dynamics marine pyroclastic volcanism: Earth and Planetary of volcanic eruptions, compaction, compression and for development and testing of predictive mod- Science Letters, v. 274, p. 40–49, doi:10.1016/j.epsl the conditions for choking: Geophysical Journal Inter- els based on eruptive precursors. .2008.06.050. national, v. 182, p. 843–864, doi:10.1111/j.1365-246X Long-term forecasting for the purpose of Ambrose, S.H., 1998, Late Pleistocene human population .2010.04674.x. bottlenecks, volcanic winter, and differentiation of Berlo, K., Blundy, J., Turner, S., Cashman, K., Hawkesworth, volcanic hazard and risk assessment requires modern humans: Journal of Human Evolution, v. 34, C., and Black, S., 2004, Geochemical precursors to detailed knowledge of past volcanic activity. p. 623–651, doi:10.1006/jhev.1998.0219. volcanic activity at Mount St. Helens, USA: Science, Volcanic hazard mapping and risk assessment Ambrose, S.H., 2003, Did the super-eruption of Toba v. 306, p. 1167–1169, doi:10.1126/science.1103869. cause a human population bottleneck? Reply to Biass, S., and Bonadonna, C., 2011, A quantitative uncer- of individual volcanic centers have evolved Gathorne-Hardy and Harcourt-Smith: Journal of tainty assessment of eruptive parameters derived from considerably over the past 25 yr, aided in large Human Evolution, v. 45, p. 231–237, doi:10.1016 tephra deposits: The example of two large eruptions of /j.jhevol.2003.08.001. Cotopaxi volcano, Ecuador: Bulletin of Volcanology, part by the development of new mapping and Andrews, B.J., and Manga, M., 2011, Effects of topography v. 73, p. 73–90, doi:10.1007/s00445-010-0404-5. computer analysis tools (such as geographic on pyroclastic density current runout and formation Bice, D.C., 1985, Quaternary volcanic stratigraphy of Mana- information systems [GIS]). Hazard maps now of coignimbrites: Geology, v. 39, p. 1099–1102, doi: gua, Nicaragua: Correlation and source assignment for 10.1130/G32226.1. multiple overlapping Plinian deposits: Geological So- exist for many volcanoes in developed countries, Annen, C., 2009, From plutons to magma chambers: ciety of America Bulletin, v. 96, no. 4, p. 553–566, doi: although there is still much work to be done in Thermal constraints on the accumulation of erupt- 10.1130/0016-7606(1985)96<553:QVSOMN>2.0.CO;2. developing countries (Aspinall et al., 2011). ible silicic magma in the upper crust: Earth and Plan- Biggs, J., Anthony, E.Y., and Ebinger, C.J., 2009, Multiple etary Science Letters, v. 284, p. 409–416, doi:10.1016 infl ation and defl ation events at Kenyan volcanoes, From a global perspective, an important ques- /j.epsl.2009.05.006. East African Rift: Geology, v. 37, p. 979–982, doi: tion relates to the frequency of very large erup- Annen, C., Blundy, J.D., and Sparks, R.S.J., 2006, The gen- 10.1130/G30133A.1. esis of intermediate and silicic magmas in deep crustal Blundy, J., and Cashman, K., 2001, Ascent-driven crystal- tions (“supervolcanoes”), that is, eruptions that hot zones: Journal of Petrology, v. 47, p. 505–539, lisation of dacite magmas at Mount St. Helens, 1980– could threaten global populations. Magnitude- doi:10.1093/petrology/egi084. 1986: Contributions to Mineralogy and Petrology, frequency relationships can be assessed using Aspinall, W.P., Auker, M., Hincks, T.K., Mahony, S.H., v. 140, p. 631–650, doi:10.1007/s004100000219. Pooley, J., Nadim, F., Syre, E., and Sparks, R.S.J., Blundy, J., and Cashman, K., 2008, Petrologic reconstruction volcano databases, such as that developed by 2011, Volcano Hazard and Exposure in Track II Coun- of magmatic system variables and processes: Reviews the Global Volcanism Program at the Smithson- tries and Risk Mitigation Measures—GFDRR Vol- in Mineralogy and Geochemistry, v. 69, p. 179–239, ian Institution (www.volcano.si.edu/index.cfm). cano Risk Study, for The World Bank, NGI Report doi:10.2138/rmg.2008.69.6. 20100806. Blundy, J., Cashman, K.V., and Berlo, K., 2008, Evolving Analysis of this database shows robust magni- Atlas, Z.D., Dixon, J.E., Sen, G., Finny, M., and Martin- Magma Storage Conditions beneath Mount St. Helens tude-frequency relationships for Holocene erup- Del Pozzo, A.L., 2006, Melt inclusions from Volcan Inferred from Chemical Variations in Melt Inclusions ≤ Popocatépetl and Volcán de Colima, Mexico: Melt from the 1980–1986 and Current (2004–2006) Erup- tions with a magnitude (M) 7 according to the evolution due to vapor-saturated crystallization dur- tions, in Sherrod, D.R., Scott, W.E., and Stauffer, P.H., Volcano Explosivity Index (Newhall and Self, ing ascent: Journal of Volcanology and Geothermal eds., A Volcano Rekindled: The Renewed Eruption of 1982), which is equivalent to the eruption of Research, v. 153, p. 221–240, doi:10.1016/j.jvolgeores Mount St. Helens, 2004–2006: U.S. Geological Survey .2005.06.010. Professional Paper 1750, 36 p. 3 ~ 100 km of magma. Larger eruptions lie off the Austin-Erickson, A., Ort, M.H., and Carrasco-Núñez, G., Bonaccorso, A., Calvari, S., Coltelli, M., Del Negro, C., and M ≤7 trend, however, which suggests that they 2011, Rhyolitic phreatomagmatism explored: Tepexitl Falsaperta, S., eds., 2004, Mt. Etna Volcano Labora- are produced by fundamentally different magma tuff ring (Eastern Mexican Volcanic Belt): Journal of tory: Geophysical Monograph 143: Washington, D.C., Volcanology and Geothermal Research, v. 201, p. 325– American Geophysical Union, 368 p. accumulation and release mechanisms than 341, doi:10.1016/j.jvolgeores.2010.09.007. Bonadonna, C., Ernst, G.G.J., and Sparks, R.S.J., 1998, smaller eruptions (Deligne et al., 2010). Concern Bachmann, O., and Bergantz, G.W., 2006, Gas percolation Thickness variations and volume estimates of tephra fall in upper-crustal silicic crystal mushes as a mechanism deposits: The importance of particle Reynolds number: about the potential impacts of a supervolcano for upward heat advection and rejuvenation of near- Journal of Volcanology and Geothermal Research, v. 81, eruption continues to spur research in this area. solidus magma bodies: Journal of Volcanology and p. 173–187, doi:10.1016/S0377-0273(98)00007-9.

Geological Society of America Bulletin, May/June 2013 683 Cashman and Sparks

Branney, M.J., and Kokelaar, P., 1992, A reappraisal of ig- of the Royal Society of London, v. 358, p. 1487–1513, ments: Geochemistry, Geophysics, Geosystems, v. 12, nimbrite emplacement: Progressive aggradation and doi:10.1098/rsta.2000.0600. Q07024, doi:10.1029/2011GC003606. changes from particulate to non-particulate fl ow during Cashman, K., and Cas, R., 2011, Introduction to the special Clague, D.A., Paduan, J.B., and Davis, A.S., 2009, Wide- emplacement of high-grade ignimbrite: Bulletin of Vol- issue of Bulletin of Volcanology, “The Cerro Galan ig- spread Strombolian eruptions of mid-ocean ridge canology, v. 54, p. 504–520, doi:10.1007/BF00301396. nimbrite and caldera: Characteristics and origins of a basalt: Journal of Volcanology and Geothermal Re- Branney, M.J., and Kokelaar, P., eds., 2003, Pyroclastic Den- very large volume ignimbrite and its magma system”: search, v. 180, p. 171–188, doi:10.1016/j.jvolgeores sity Currents and the Sedimentation of Ignimbrites: Bulletin of Volcanology, v. 73, no. 10, p. 1425–1426, .2008.08.007. Geological Society of London Memoir 27, 143 p. doi:10.1007/s00445-011-0559-8. Clarke, A.B., Voight, B., Neri, A., and Macedonio, G., Brown, L.D., Zhao, W.J., Nelson, D.K., Hauck, M., Alsdorf, Cashman, K.V., 1992, Groundmass crystallization of Mount 2002, Transient dynamics of Vulcanian explosions D., and Ross, A., 1996, Bright spots, structure, and St. Helens dacite, 1980–1986—A tool for interpreting and column collapse: Nature, v. 415, p. 897–901, doi: magmatism in southern Tibet from INDEPTH seismic shallow magmatic processes: Contributions to Miner- 10.1038/415897a. refl ection profi ling: Science, v. 274, p. 1688–1690, alogy and Petrology, v. 109, p. 431–449, doi:10.1007 Coffey, T.S., 2008, Diet Coke and Mentos: What is really doi:10.1126/science.274.5293.1688. /BF00306547. behind this physical reaction?: American Journal of Brown, M., 2007, Crustal melting and melt extraction, ascent Cashman, K.V., 2004, Volatile controls on magma ascent Physics, v. 76, p. 551–557, doi:10.1119/1.2888546. and emplacement in orogens: Mechanisms and conse- and eruption, in Sparks, R.S.J., and Hawkesworth, C.J., Coltelli, M., Del Carlo, P., and Vezzoli, L., 1998, The discovery quences: Journal of the Geological Society of London, The State of the Planet: Frontiers and Challenges in of a Plinian basaltic eruption of Roman age at Etna vol- v. 164, p. 709–730, doi:10.1144/0016-76492006-171. Geophysics: Geophysical Monograph 150: Washing- cano, Italy: Geology, v. 26, p. 1095–1098, doi:10.1130 Brown, R.J., Kavanagh, J., Sparks, R.S.J., Tait, M., and Field, ton, D.C., American Geophysical Union, p. 109–124, /0091-7613(1998)026<1095:DOAPBE>2.3.CO;2. M., 2007, Mechanically disrupted and chemically weak- doi:10.1029/150GM10. Condomines, M., Gauthier, P.-J., and Sigmarsson, O., 2003, ened zones in segmented dike systems cause vent lo- Cashman, K.V., and Fiske, R.S., 1991, Fallout of pyro- Timescales of magma chamber processes and dating of calization: Evidence from kimberlite volcanic systems: clastic debris from submarine volcanic eruptions: young volcanic rocks: Reviews in Mineralogy and Geo- Geology, v. 35, p. 815–818, doi:10.1130/G23670A.1. Science, v. 253, p. 275–280, doi:10.1126/science.253 chemistry, v. 52, p. 125–174, doi:10.2113/0520125. Bruce, P.M., and Huppert, H.E., 1989, Thermal control of .5017.275. Cooper, K.M., and Reid, M.R., 2008, Uranium-series crystal basaltic fi ssure eruptions: Nature, v. 342, p. 665–667, Cashman, K.V., and Giordano, G., 2008, Volcanoes and ages: Reviews in Mineralogy and Geochemistry, v. 69, doi:10.1038/342665a0. human history: Journal of Volcanology and Geo- p. 479–544, doi:10.2138/rmg.2008.69.13. Burden, R.E., Phillips, J.C., and Hincks, T.K., 2011, Esti- thermal Research, v. 176, p. 325–329, doi:10.1016 Costa, A., Melnik, O., Sparks, R.S.J., and Voight, B., 2006, mating volcanic plume heights from depositional clast /j.jvolgeores.2008.01.036. The control of magma fl ow in dykes on cyclic lava size: Journal of Geophysical Research, v. 116, B11206, Cashman, K.V., and McConnell, S., 2005, Transitions from dome extrusion: Geophysical Research Letters, v. 34, doi:10.1029/2011JB008548. explosive to effusive activity—The summer 1980 erup- L02303, doi:10.1029/2006GL027466. Burgisser, A., Arbaret, L., Druitt, T.H., and Giachetti, T., tions of Mount St. Helens: Bulletin of Volcanology, Costa, A., Caricchi, L., and Bagdassarov, N., 2009, A model 2011, Pre-explosive conduit conditions of the 1997 v. 68, p. 57–75, doi:10.1007/s00445-005-0422-x. for the rheology of particle-bearing suspensions and Vulcanian explosions at Soufrière Hills Volcano, Mont- Cashman, K.V., Thornber, C.R., and Kauahikaua, J.P., 1999, partially molten rocks: Geochemistry, Geophysics, serrat: II. Overpressure and depth distributions: Jour- Cooling and crystallization of lava in open channels, Geosystems, v. 10, 13 p. doi:10.1029/2008GC002138. nal of Volcanology and Geothermal Research, v. 199, and the transition of pahoehoe lava to ‘a‘ā: Bulle- Costa, F., and Morgan, D., 2010, Time constraints from chem- p. 193–205, doi:10.1016/j.jvolgeores.2010.11.014. tin of Volcanology, v. 61, p. 306–323, doi:10.1007 ical equilibration in magmatic crystals, in Dosseto, A., Bursik, M., 2001, The effect of wind on the rise height of /s004450050299. Turner, S., and Van-Orman, J., eds., Timescales of Mag- volcanic plumes: Geophysical Research Letters, v. 28, Cashman, K.V., Kerr, R.C., and Griffi ths, R.W., 2006, A lab- matic Processes: Chichester, UK, John Wiley & Sons, p. 3621, doi:10.1029/2001GL013393. oratory model of surface crust formation and disruption Ltd., p. 125–159. Büttner, R., Dellino, P., and Zimanowski, B., 1999, Identifying on channelized lava fl ows: Bulletin of Volcanology, Costantini, L., Houghton, B.F., and Bonadonna, C., 2010, Con- magma-water interaction from the surface features of ash v. 68, p. 753–770, doi:10.1007/s00445-005-0048-z. straints on eruption dynamics of basaltic explosive activity particles: Nature, v. 401, p. 688–690, doi:10.1038/44364. Cashman, K.V., Thornber, C.R., and Pallister, J.S., 2008, derived from chemical and microtextural study: The ex- Cabrera, A., Weinberg, R.F., Wright, H.M.N., Zlotnik, S., From Dome to Crystallization and Fragmentation of ample of the Fontana Lapilli Plinian eruption, Nicaragua: and Cas, R.A.F., 2011, Melt fracturing and healing: A Conduit Magma during the 2004–2006 Dome Extru- Journal of Volcanology and Geothermal Research, v. 189, mechanism for degassing and origin of silicic obsidian: sion of Mount St. Helens, in Sherrod, D.R., Scott, p. 207–224, doi:10.1016/j.jvolgeores.2009.11.008. Geology, v. 39, p. 67–70, doi:10.1130/G31355.1. W.E., and Stauffer, P.H., eds., A Volcano Rekindled: Couch, S., Sparks, R.S.J., and Carroll, M.R., 2001, Min- Calder, E.S., Cole, P.D., Dade, W.B., Druitt, T.H., Hoblitt, The Renewed Eruption of Mount St. Helens, 2004– eral disequilibrium in lavas explained by convective R.P., Huppert, H.E., Ritchie, L., Sparks, R.S.J., and 2006: U.S. Geological Survey Professional Paper self-mixing in open magma chambers: Nature, v. 411, Young, S.R., 1999, Mobility of pyroclastic fl ows and 1750, p. 387–413. p. 1037–1039, doi:10.1038/35082540. surges at the Soufriere Hills volcano, Montserrat: Castro, J.M., Manga, M., and Cashman, K.V., 2002, Dynam- Daniels, K., Kavanagh, J., Menand, T., and Sparks, R.S.J., Geophysical Research Letters, v. 26, p. 537–540, doi: ics of obsidian fl ows inferred from microstructures: 2012, The shapes of dykes: Evidence for the infl u- 10.1029/1999GL900051. Insights from microlite preferred orientations: Earth ence of cooling and inelastic deformation: Geological Calvache, M.L., and Williams, S.N., 1997, Emplacement and Planetary Science Letters, v. 199, p. 211–226, doi: Society of America Bulletin, v. 124, p. 1102–1112, and petrological evolution of the andesitic dome of 10.1016/S0012-821X(02)00559-9. doi:10.1130/B30537.1. Galeras volcano, 1990–1992: Journal of Volcanology Castruccio, A., Rust, A., and Sparks, R.S.J., 2010, Rheology Davis, A., and Clague, D., 2006, Volcaniclastic deposits and Geothermal Research, v. 77, p. 57–69, doi:10.1016 and fl ow of crystal-rich bearing lavas: Insights from ana- from the North Arch volcanic fi eld, Hawaii: Explosive /S0377-0273(96)00086-8. logue gravity currents: Earth and Planetary Science Let- fragmentation of alkalic lava at abyssal depths: Bul- Calvari, S., Inguggiato, S., Puglisi, G., Ripepe, M., and Rosi, ters, v. 297, p. 471–480, doi:10.1016/j.epsl.2010.06.051. letin of Volcanology, v. 68, p. 294–307, doi:10.1007 M., eds., 2008, The Stromboli Volcano: An integrated Chadwick, W.W., Jr., Cashman, K.V., Embley, R.W., Matsu- /s00445-005-0008-7. study of the 2002–2003 eruption: Geophysical Mono- moto, H., Dziak, R.P., de Ronde, C.E.J., Lau, T.K., Deardorff, N.D., Cashman, K.V., and Chadwick, W.W., Jr., graph 182: Washington, D.C., American Geophysical Deardorff, N.D., and Merle, S.G., 2008, Direct video 2011, Observations of eruptive plume dynamics and py- Union, doi:10.1029/GM182, 399 p. and hydrophone observations of submarine explo- roclastic deposits from submarine explosive eruptions Caress, D.W., Clague, D.A., Paduan, J.B., Martin, J.F., sive eruptions at NW Rota-1 volcano, Mariana arc: at NW Rota-1, Mariana arc: Journal of Vol canol ogy and Dreyer, B.M., Chadwick, W.W., Denny, A., and Kel- Journal of Geophysical Research, v. 113, B08S10, Geothermal Research, v. 202, p. 47–59, doi:10.1016 ley, D.S., 2012, Repeat bathymetric surveys at 1-metre doi:10.1029/2007JB005215. /j.jvolgeores.2011.01.003. resolution of lava fl ows erupted at Axial Seamount Chadwick, W.W., Jr., Nooner, S.L., Butterfi eld, D.A., and Deligne, N.I., Coles, S.G., and Sparks, R.S.J., 2010, Re- in April 2011: Nature Geoscience, v. 5, p. 483–488, Lilley, M.D., 2012, Seafl oor deformation and forecasts currence rates of large explosive volcanic eruptions: doi:10.1038/ngeo1496. of the April 2011 eruption at Axial Seamount: Nature Journal of Geophysical Research, v. 115, B06203, Carey, S., and Sparks, R.S.J., 1986, Quantitative models of Geoscience, v. 5, p. 474–477, doi:10.1038/ngeo1464. doi:10.1016/j.jvolgeores.2008.01.036. the fallout and dispersal of tephra from volcanic erup- Chaput, J.A., Zandomeneghi, D., and Aster, R.C., 2012, Dellino, P., Dioguardi, F., Zimanowski, B., Buttner, R., Mele, tion columns: Bulletin of Volcanology, v. 48, p. 109– Imaging of Erebus volcano using body wave seismic D., La Volpe, L., Sulpizio, R., Doronzo, D.M., Sonder, 125, doi:10.1007/BF01046546. interferometry of Strombolian eruption coda: Geo- I., Bonasia, R., Calvari, S., and Marotta, E., 2010, Carn, S.A., Pallister, J.S., Lara, L., Ewert, J.W., Watt, S., physical Research Letters, v. 39, L07304, doi:10.1029 Conduit fl ow experiments help constrain the regime of Prata, A.J., Thomas, R.J., and Villarosa, G., 2009, The /2012GL050956. explosive eruptions: Journal of Geophysical Research, unexpected awakening of Chaiten volcano, Chile: Eos Charlier, B.L.A., Wilson, C.J.N., Lowenstern, J.B., Blake, v. 115, no. B4, B04204, doi:10.1029/2009JB006781. (Transactions, American Geophysical Union), v. 90, S., Van Calsteren, P., and Davidson, J.P., 2005, Magma Dellino, P., Gudmundsson, M.T., Larsen, G., Mele, D., no. 24, p. 205, doi:10.1029/2009EO240001. generation at a large hyperactive silicic volcano Stevenson, J.A., Thordarson, T., and Zimanowski, B., Cas, R.A.F., and Giordano, G., 2006, Explosive mafi c vol- (Taupo, New Zealand) revealed by U-Th and U-Pb 2012, Ash from the Eyjafjallajökull eruption (Iceland): canism—Preface: Journal of Volcanology and Geo- systematics in zircons: Journal of Petrology, v. 46, Fragmentation processes and aerodynamic behavior: thermal Research, v. 155, p. vi, doi:10.1016/j.jvolgeores p. 3–32, doi:10.1093/petrology/egh060. Journal of Geophysical Research, v. 117, B00C04, .2006.02.003. Cimarelli, C., Costa, A., Mueller, S., and Mader, H.M., doi:10.1029/2011JB008726. Cashman, K., and Blundy, J., 2000, Degassing and crystalliza- 2011, Rheology of magmas with bimodal crystal size De Silva, S., Zandt, G., Trumbull, R., Viramonte, J.G., tion of ascending andesite: Philosophical Transactions and shape distributions: Insights from analog experi- Salas, G., and Jimenez, N., 2006, Large ignimbrite

684 Geological Society of America Bulletin, May/June 2013 How volcanoes work

eruptions and volcano-tectonic depressions in the p. 603–606, doi:10.1130/0091-7613(2000)28<603: ters, v. 168, p. 201–218, doi:10.1016/S0012-821X Central Andes: A thermomechanical perspective, in MICRCZ>2.0.CO;2. (99)00051-5. Troise, C., De Natale , G., and Kilburn, C.R.J., eds., Embley, R.W., and 16 others, 2006, Long-term eruptive Garry, W.B., and Bleacher, J.E., eds., 2011, Analogs for Mechanisms of Activity and Unrest at Large Calderas: activity at a submarine arc volcano: Nature, v. 441, Planetary Exploration: Geological Society of America Geological Society of London Special Publication 269, p. 494–497, doi:10.1038/nature04762. Special Paper 483, 567 p. p. 47–63. Ernst, G.G.J., Sparks, R.S.J., Carey, S.N., and Burisk, M.I., Geschwind, C.-H., and Rutherford, M.J., 1995, Crystalliza- Dietterich, H.R., Poland, M.P., Schmidt, D.A., Cashman, 1996, Sedimentation from turbulent jets and plumes: tion of microlites during magma ascent: The fl uid me- K.V., Sherrod, D.R., and Espinosa, A.T., 2012, Track- Journal of Geophysical Research, v. 101, no. B3, chanics of 1980–1986 eruptions at Mount St Helens: ing lava fl ow emplacement on the east rift zone of p. 5575–5589, doi:10.1029/95JB01900. Bulletin of Volcanology, v. 57, p. 356–370. Kilauea, Hawai’i, with synthetic aperture radar (SAR) Favalli, M., Fornaciai, A., Mazzarini, F., Harris, A., Neri, Giordano, D., and Dingwell, D.B., 2003, Non-Arrhenian coherence: Geochemistry, Geophysics, Geosystems, M., Behncke, B., Pareschi, M.T., Tarquini, S., and muticomponent melt viscosity: Earth and Planetary v. 13, Q05001, doi:10.1029/2011GC004016. Boschi, E., 2010, Evolution of an active lava fl ow Science Letters, v. 208, p. 337–349. Dingwell, D.B., 1998, The glass transition in hydrous gra- fi eld using a multitemporal LIDAR acquisition: Giordano, D., Russell, J.K., and Dingwell, D.B., 2008, Vis- nitic melts: Physics of the Earth and Planetary Interiors, Journal of Geophysical Research, v. 115, B11203, cosity of magmatic liquids: A model: Earth and Plan- v. 107, p. 1–8, doi:10.1016/S0031-9201(97)00119-2. doi:10.1029/2010JB007463. etary Science Letters, v. 271, p. 123–134, doi:10.1016 Dingwell, D.B., and Webb, S.L., 1989, Structural relaxation Fedele, F.G., Giaccio, B., and Hajdas, I., 2008, Timescales /j.epsl.2008.03.038. in silicate melts and non-Newtonian melt rheology in and cultural process at 40,000 yrs. BP in the light of Giordano, G., 1998, The effect of paleotopography on lithic geologic processes: Physics and Chemistry of Miner- the Campanian Ignimbrite eruption, western Eurasia: distribution and facies associations of small volume als, v. 16, p. 508–516, doi:10.1007/BF00197020. Journal of Human Evolution, v. 55, no. 5, p. 834–857, ignimbrites: The WTT Cupa (Roccamonfi na volcano, Dingwell, D.B., Romano, C., and Hess, K.-U., 1996, The doi:10.1016/j.jhevol.2008.08.012. Italy): Journal of Volcanology and Geothermal Re- effect of water on the viscosity of a haplogranitic melt Fialko, Y., 2001, On origin of near-axis volcanism and fault- search, v. 87, no. 1–4, p. 255–273, doi:10.1016/S0377 under P-T-X conditions relevant to silicic volcanism: ing at fast spreading mid-ocean ridges: Earth and Plan- -0273(98)00096-1. Contributions to Mineralogy and Petrology, v. 124, etary Science Letters, v. 190, p. 31–39, doi:10.1016 Girolami, L., Roche, O., Druitt, T.H., and Corpetti, T., 2010, p. 19–28, doi:10.1007/s004100050170. /S0012-821X(01)00376-4. Particle velocity fi elds and depositional processes in Di Stefano, R., and Chiarabba, C., 2002, Active source Fierstein, J., and Nathenson, M., 1992, Another look at the laboratory ash fl ows, with implications for the sedi- tomog raphy at Mt. Vesuvius: Constraints for the mag- calculation of fallout tephra volumes: Bulletin of Vol- mentation of dense pyroclastic fl ows: Bulletin of Vol- matic system: Journal of Geophysical Research, v. 107, canology, v. 54, p. 156–167, doi:10.1007/BF00278005. canology, v. 72, p. 747–759, doi:10.1007/s00445-010 no. B11, p. 2278, doi:10.1029/2001JB000792. Fierstein, J., and Wilson, C.J.N., 2005, Assembling an ig- -0356-9. Doyle, E.E., Hogg, A.J., Mader, H.M., and Sparks, R.S.J., nimbrite: Compositionally defi ned eruptive packages Gonnermann, H., and Manga, M., 2003, Explosive vol- 2008, Modeling dense pyroclastic basal flows from in the 1912 Valley of Ten Thousand Smokes ignim- canism may not be an inevitable consequence of collapsing columns: Geophysical Research Letters, brite, Alaska: Geological Society of America Bulletin, magma fragmentation: Nature, v. 426, p. 432–435, v. 35, L04305, doi:10.1029/2007GL032585. v. 117, p. 1094–1107, doi:10.1130/B25621.1. doi:10.1038/nature02138. Druitt, T.H., 1992, Emplacement of the 18 May 1980 lat- Fiske, R.S., Naka, J., Iizasa, K., Yuasa, M., and Klaus, A., Grattan, J., and Torrence, R., 2010, Living Under the eral blast deposit ENE of Mount St. Helens, Wash- 2001, Submarine silicic caldera at the front of the Shadow: Cultural Impacts of Volcanic Eruptions: Wal- ington: Bulletin of Volcanology, v. 54, p. 554–572, Izu-Bonin Arc, Japan: Voluminous seafl oor eruptions nut Creek, California, Left Coast Press, 320 p. doi:10.1007/BF00569940. of rhyolite pumice: Geological Society of America Gravley, D.M., Wilson, C.J.N., Leonard, G.S., and Cole, Druitt, T.H., 1998, Pyroclastic density currents, in Gilbert, Bulletin, v. 113, p. 813–824, doi:10.1130/0016-7606 P.W., 2007, Double trouble: Paired ignimbrite erup- J.S., and Sparks, R.S.J., eds., The Physics of Explo- (2001)113<0813:SSCATF>2.0.CO;2. tions and collateral subsidence in the Taupo volcanic sive Volcanic Eruptions: Geological Society of Lon- Fiske, R.S., Rose, T.R., Swanson, D.A., Champion, D.E., zone, New Zealand: Geological Society of America don Special Publication 145, p. 145–182, doi:10.1144 and McGeehin, J.P., 2009, Kulanaokuaiki Tephra (ca. Bulletin, v. 119, p. 18–30, doi:10.1130/B25924.1. /GSL.SP.1996.145.01.08. A.D. 400–1000): Newly recognized evidence for highly Griffi ths, R.W., 2000, The dynamics of lava fl ows: An- Druitt, T.H., Mellors, R.A., Pyle, D.M., and Sparks, R.S.J., explosive eruptions at Kīlauea volcano, Hawai‘i: Geo- nual Review of Fluid Mechanics, v. 32, p. 477–518, 1989, Explosive volcanism on Santorini, Greece: logical Society of America Bulletin, v. 121, p. 712–728, doi:10.1146/annurev.fl uid.32.1.477. Geological Magazine, v. 126, p. 95–126, doi:10.1017 doi:10.1130/B26327.1. Griffi ths, R.W., Kerr, R.C., and Cashman, K.V., 2003, Pat- /S0016756800006270. Fournier, T.J., Pritchard, M.E., and Riddick, S.N., 2010, Du- terns of solidifi cation in channel fl ows with surface Druitt, T.H., Calder, E.S., Cole, P.D., Hoblitt, R.P., Lough- ration, magnitude, and frequency of subaerial volcano cooling: Journal of Fluid Mechanics, v. 496, p. 33–62, lin, S.C., Norton, G.E., Ritchie, L.J., Sparks, R.S.J., deformation events: New results from Latin America doi:10.1017/S0022112003006517. and Voight, B., 2002a, Small-volume, highly mobile using InSAR and a global synthesis: Geochemistry, Gudmundsson, A., 1987, Lateral magma fl ow, caldera col- pyroclastic fl ows formed by rapid sedimentation from Geophysics, Geosystems, v. 11, no. 1, p. Q01003, lapse, and a mechanism of large eruptions in Iceland: pyroclastic surges at Soufrière Hills volcano, Mont- doi:10.1029/2009GC002558. Journal of Volcanology and Geothermal Research, serrat: An important volcanic hazard: Geological So- Fowler, A.C., Scheu, B., Lee, W.T., and McGuinness, M.J., v. 34, p. 65–78, doi:10.1016/0377-0273(87)90093-X. ciety of London Memoir 21, p. 263–279. 2010, A theoretical model of the explosive fragmen- Gudmundsson, A., 1988, Formation of collapse calderas: Druitt, T.H., Young, S.R., Baptie, B., Bonadonna, C., Calder, tation of vesicular magma: Proceedings of the Royal Geology, v. 16, p. 808–810, doi:10.1130/0091-7613 E.S., Clarke, A.B., Cole, P.D., Harford, C.L., Herd, Society A: Mathematical, Physical and Engineering (1988)016<0808:FOCC>2.3.CO;2. R.A., Luckett, R., Ryan, G., and Voight, B., 2002b, Science, v. 466, p. 731–752. Gurioli, L., Houghton, B.F., Cashman, K.V., and Cioni, Episodes of cyclic Vulcanian explosive activity with Fox, C.G., Chadwick, W.W., and Embley, R.W., 2001, Direct R., 2005, Complex changes in eruption dynamics dur- fountain collapse at Soufrière Hills volcano, Mont- observation of a submarine volcanic eruption from a sea- ing the 79 AD eruption of Vesuvius: Bulletin of Vol- serrat, in Druitt, T.H., and Kokellar, B.P., eds., The fl oor instrument caught in a lava fl ow: Nature, v. 412, canology, v. 67, p. 144–159, doi:10.1007/s00445 Eruption of Soufriere Hills Volcano, Montserrat from p. 727–729, doi:10.1038/35089066. -004-0368-4. 1995 to 1999: Geological Society of London Memoir Francis, P., Glaze, L., and Oppenheimer, C., 1990, Eruption Hamada, M., Laporte, D., Cluzel, N., Koga, K., and Kawa- 21, p. 281–306. terms: Nature, v. 346, p. 519, doi:10.1038/346519a0. moto, T., 2010, Simulating bubble number density of Druitt, T.H., Avard, G., Bruni, G., Lettieri, P., and Maez, Frey, H., and Lange, R., 2011, Phenocryst complexity in rhyolitic pumices from Plinian eruptions: Constraints F., 2007, Gas retention in fi ne-grained pyro clastic andesites and dacites from the Tequila volcanic fi eld, from fast decompression experiments: Bulletin of Vol- flow materials at high temperatures: Bulletin of Mexico: Resolving the effects of degassing vs. magma canology, v. 72, p. 735–746, doi:10.1007/s00445-010 Volcanology, v. 69, no. 8, p. 881–901, doi:10.1007 mixing: Contributions to Mineralogy and Petrology, -0353-z. /s00445-007-0116-7. v. 162, p. 415–445, doi:10.1007/s00410-010-0604-1. Hammer, J.E., 2008, Experimental studies of the kinetics Druitt, T.H., Costa, F., Deloule, E., Dungan, M., and Scaillet, Funiciello, R., and Giordano, G., eds., 2010, The Colli Al- and energetics of magma crystallization: Reviews B., 2012, Decadal to monthly timescales of magma trans- bani Volcano: Special Publications of International in Mineralogy and Geochemistry, v. 69, p. 9–59, doi: fer and reservoir growth at a caldera volcano: Nature, Asso ciation of Volcanology and Chemistry of the 10.2138/rmg.2008.69.2. v. 482, p. 77–80, doi:10.1038/nature10706. Earth’s Interior, Volume 3: London, Geological Society Hammer, J.E., and Rutherford, M.J., 2002, An experimental Dziak, R.P., Haxel, J.H., Bohnenstiehl, D.R., Chadwick, of London, 276 p. study of the kinetics of decompression-induced crys- W.W., Nooner, S.L., Fowler, M.J., Matsumoto, H., Gaillard, J.C., and Texier, P., 2010, Religions, natural haz- tallization in silicic melt: Journal of Geophysical Re- and Butterfi eld, D.A., 2012, Seismic precursors and ards, and disasters: An introduction: Religion, v. 40, search, v. 107, 2021, doi:10.1029/2001JB000281. magma ascent before the April 2011 eruption at Axial p. 81–84, doi:10.1016/j.religion.2009.12.001. Hammer, J.E., and Rutherford, M.J., 2003, Petrologic indica- Seamount: Nature Geoscience, v. 5, p. 478–482, doi: Gardner, J.E., Thomas, R.M.E., Jaupart, C., and Tait, S., tors of pre-eruption magma dynamics: Geology, v. 31, 10.1038/ngeo1490. 1996, Fragmentation of magma during Plinian volcanic p. 79–82, doi:10.1130/0091-7613(2003)031<0079: Eichelberger, J.C., Carrigan, C.R., Westrich, H.R., and Price, eruptions: Bulletin of Volcanology, v. 58, p. 144–162, PIOPMD>2.0.CO;2. R.H., 1986, Non-explosive silicic volcanism: Nature, doi:10.1007/s004450050132 Hammer, J.E., Cashman, K.V., and Voight, B., 2000, Mag- v. 323, p. 598–602, doi:10.1038/323598a0. Gardner, J.E., Hilton, M., and Carroll, M.R., 1999, Experi- matic processes revealed by textural and compositional Eichelberger, J.C., Chertkoff, D.G., Dreher, S.T., and mental constraints on degassing of magma: Isother- trends in Merapi dome lavas: Journal of Volcanology Nye, C.J., 2000, Magmas in collision: Rethinking mal bubble growth during continuous decompression and Geothermal Research, v. 100, p. 165–192, doi: chemical zonation in silicic magmas: Geology, v. 28, from high pressure: Earth and Planetary Science Let- 10.1016/S0377-0273(00)00136-0.

Geological Society of America Bulletin, May/June 2013 685 Cashman and Sparks

Hansell, A., and Oppenheimer, C., 2004, Health hazards Hawai‘i: U.S. Geological Survey Professional Paper cano, Hawai‘i: The First 20 Years: U.S. Geological from volcanic gases: A systematic literature review: 1676, p. 89–104. Survey Professional Paper 1676, p. 63–88. Archives of Environmental Health: An International Horwell, C., and Baxter, P., 2006, The respiratory health Kavanagh, J.L., and Sparks, R.S.J., 2011, Insights of dyke Journal (Toronto, Ontario), v. 59, p. 628–639. hazards of volcanic ash: A review for volcanic risk emplacement mechanics from detailed 3D dyke thick- Harford, C.L., Sparks, R.S.J., and Fallick, A.E., 2003, De- mitigation: Bulletin of Volcanology, v. 69, p. 1–24, doi: ness datasets: Journal of the Geological Society, v. 168, gassing at the Soufrière Hills volcano, Mont serrat, 10.1007/s00445-006-0052-y. p. 965–978, doi:10.1144/0016-76492010-137. recorded in matrix glass compositions: Journal of Houghton, B.F., and Wilson, C.J.N., 1989, A vesicularity Kavanagh, J., Menand, T., and Sparks, R.S.J., 2006, An experi- Petrol ogy, v. 44, p. 1503–1523, doi:10.1093/petrology index for pyroclastic deposits: Bulletin of Volcanology, mental investigation of sill formation and propagation in /44.8.1503. v. 51, no. 6, p. 451–462, doi:10.1007/BF01078811. layered elastic media: Earth and Planetary Science Let- Harrington, R.M., and Brodsky, E.E., 2007, Volcanic hybrid Hulme, G., 1974, The interpretation of lava fl ow morphology: ters, v. 245, p. 799–813, doi:10.1016/j.epsl.2006.03.025. earthquakes that are brittle-failure events: Geophysi- Geophysical Journal of the Royal Astronomical Soci- Kent, A.J.R., Darr, C., Koleszar, A.M., Salisbury, M.J., and cal Research Letters, v. 34, L06308, doi:10.1029 ety, v. 39, p. 361–383, doi:10.1111/j.1365-246X.1974 Cooper, K.M., 2010, Preferential eruption of andesitic /2006GL028714. .tb05460.x. magmas through recharge fi ltering: Nature Geoscience, Harris, A., and Ripepe, M., 2007, Synergy of multiple geo- Humphreys, M., Blundy, J.D., and Sparks, R.S.J., 2006, v. 3, p. 631–636, doi:10.1038/ngeo924. physical approaches to unravel explosive eruption Magma evolution and open-system processes at Shive- Kerr, R.C., 2001, Thermal erosion by laminar lava conduit and source dynamics—A case study from luch volcano; insights from phenocryst zoning: Jour- flows: Journal of Geophysical Research, v. 106, Stromboli: Chemie der Erde–Geochemistry, v. 67, nal of Petrology, v. 47, p. 2303–2334, doi:10.1093 p. 26,453–26,465. p. 1–35. /petrology/egl045. Kerr, R.C., and Lister, J.R., 1991, The effects of shape on crys- Harris, A., and Rowland, S.R., 2001, FLOWGO: A kine- Huppert, H.E., and Woods, A.W., 2002, The role of volatiles tal settling and on the rheology of magmas: The Journal matic thermo-rheological model for lava fl owing in a in magma chamber dynamics: Nature, v. 420, p. 493– of Geology, v. 99, p. 457–467, doi:10.1086/629506. channel: Bulletin of Volcanology, v. 63, p. 20–44, doi: 495, doi:10.1038/nature01211. Kerr, R.C., and Lyman, A.W., 2007, Importance of sur- 10.1007/s004450000120. Hurwitz, S., and Navon, O., 1994, Bubble nucleation in rhyo- face crust strength during the fl ow of the 1988–1990 Harris, A., Dehn, J., Patrick, M., Calvari, S., Ripepe, M., and litic melts: Experiments at high pressure, temperature, andesite lava of Lonquimay volcano, Chile: Journal Lodato, L., 2005, Lava effusion rates from hand-held and water content: Earth and Planetary Science Letters, of Geophysical Research, v. 112, no. B3, B03209, thermal infrared imagery: An example from the June v. 122, p. 267–280, doi:10.1016/0012-821X(94)90001-9. doi:10.1029/2006JB004522. 2003 effusive activity at Stromboli: Bulletin of Volcanol- Iizasa, K., Fiske, R.S., Ishizuka, O., Yuasa, M., Hashimoto, Kerr, R.C., Griffi ths, R.W., and Cashman, K.V., 2006, ogy, v. 68, p. 107–117, doi:10.1007/s00445-005-0425-7. J., Ishibashi, J., Naka, J., Horii, Y., Fujiwara, Y., Imai, Formation of channelized lava fl ows on an uncon- Harris, A., Dehn, J., and Calvari, S., 2007, Lava effusion A., and Koyama, S., 1999, A Kuroko-type polymetallic fi ned slope: Journal of Geophysical Research, v. 111, rate defi nition and measurement: A review: Bulletin of sulfi de deposit in a submarine silicic caldera: Science, B10206, doi:10.1029/2005JB004225. Vol canology, v. 70, p. 1–22, doi:10.1007/s00445-007 v. 283, p. 975–977. Kilburn, C.R.J., 2004, Fracturing as a quantitative indica- -0120-y. James, M., Lane, S.J., Wilson, L., and Corder, S.D., 2009, tor of lava fl ow dynamics: Journal of Volcanology and Hawkesworth, C., George, R., Turner, S., and Zellmer, G., Degassing at low magma-viscosity volcanoes: Quan- Geothermal Research, v. 132, p. 209–224, doi:10.1016 2004, Time scales of magmatic processes: Earth and tifying the transition between passive bubble-burst /S0377-0273(03)00346-9. Planetary Science Letters, v. 218, p. 1–16, doi:10.1016 and Strombolian eruption: Journal of Volcanology and Klug, C., and Cashman, K.V., 1996, Permeability relation- /S0012-821X(03)00634-4. Geothermal Research, v. 180, p. 81–88, doi:10.1016 ships in vesiculating magmas: Implications for frag- Head, J.W., III, and Wilson, L., 2003, Deep submarine pyro- /j.jvolgeores.2008.09.002. mentation: Bulletin of Volcanology, v. 58, p. 87–100, clastic eruptions: Theory and predicted landforms James, M., Pinkerton, H., and Ripepe, M., 2010, Imaging short doi:10.1007/s004450050128. and deposits: Journal of Volcanology and Geother- period variations in lava fl ux: Bulletin of Volcanology, Kokelaar, P., 1986, Magma-water interactions in subaqueous mal Research, v. 121, p. 155–193, doi:10.1016/S0377 v. 72, p. 671–676, doi:10.1007/s00445-010-0354-y. and emergent basaltic volcanism: Bulletin of Volcanol- -0273(02)00425-0. Jaupart, C., and Allegre, C.J., 1991, Gas content, eruption ogy, v. 48, p. 275–289, doi:10.1007/BF01081756. Heliker, C.C., Swanson, D.A., and Takahashi, T.J., eds., rate and instabilities of eruption regime in silicic vol- Koulakov, I., Gordeev, E.I., Dobretsov, N.L., Vernikovsky, 2003, The Pu‘u ‘Ō‘ō-Kūpaianaha Eruption of Kīlauea canoes: Earth and Planetary Science Letters, v. 102, V.A., Senyukov, S., and Jakovlev, A., 2011, Feeding Volcano, Hawai‘i: The First 20 Years: U.S. Geological p. 413–429, doi:10.1016/0012-821X(91)90032-D. volcanoes of the Kluchevskoy group from the results of Survey Professional Paper 1676, 215 p. Jellinek, A.M., and DePaolo, D.J., 2003, A model for the local earthquake tomography: Geophysical Research Helz, R.T., Banks, N.G., Heliker, C., Neal, C.A., and Wolfe, origin of large silicic magma chambers: Precursors of Letters, v. 38, p. L09305, doi:10.1029/2011GL046957. E.W., 1995, Comparative geothermometry of recent caldera-forming eruptions: Bulletin of Volcanology, Koyaguchi, T., and Woods, A.W., 1996, On the formation of Hawaiian eruptions: Journal of Geophysical Research, v. 65, p. 363–381, doi:10.1007/s00445-003-0277-y. eruption columns following explosive mixing of magma v. 100, p. 17,637–17,657, doi:10.1029/95JB01309. Kaminski, E., and Jaupart, C., 1997, Expansion and quench- and surface-water: Journal of Geophysical Research, Helz, R.T., Heliker, C., Hon, K., and Mangan, M., 2003, ing of vesicular magma fragments in Plinian eruptions: v. 101, no. B3, p. 5561–5574, doi:10.1029/95JB01687. Thermal effi ciency of lava tubes in the Pu‘u ‘Ō‘ō- Journal of Geophysical Research, v. 102, p. 12,187– Koyaguchi, T., Scheu, B., Mitani, N.K., and Melnik, O., 2008, Kūpaianaha eruption, in Heliker, C.C., Swanson, D.A., 12,203, doi:10.1029/97JB00622. A fragmentation criterion for highly viscous bubbly and Takahashi, T.J., eds., The Pu‘u ‘Ō‘ō-Kūpaianaha Kano, K., 2003, Subaqueous pumice eruptions and their prod- magmas estimated from shock tube experiments: Jour- Eruption of Kīlauea Volcano, Hawai‘i: The First 20 ucts, in White, J.D.L., Smellie, J.L., and Clague, D.A., nal of Volcanology and Geothermal Research, v. 178, Years: U.S. Geological Survey Professional Paper eds., Explosive Subaqueous Volcanism: Geophysical p. 58–71, doi:10.1016/j.jvolgeores.2008.02.008. 1676, p. 105–120. Monograph 140: Washington, D.C., American Geophysi- Kueppers, U., Scheu, B., Spieler, O., and Dingwell, D.B., Hildreth, W.S., 2006, Volcanological perspectives on Long cal Union, p. 213–229, doi:10.1029/140GM14. 2006a, Fragmentation effi ciency of explosive volcanic Valley, Mammoth Mountains, and Mono Craters: Sev- Katz, M.G., and Cashman, K.V., 2003, Hawaiian lava fl ows eruptions: A study of experimentally generated pyro- eral contiguous but discrete systems: Journal of Vol- in the third dimension: identifi cation and interpreta- clasts: Journal of Volcanology and Geothermal Research, canology and Geothermal Research, v. 13, p. 169–198. tion of pahoehoe and ‘a‘a distribution in the KP-1 and v. 153, p. 125–135, doi:10.1016/j.jvolgeores.2005 Ho, A.M., and Cashman, K.V., 1997, Temperature con- SOH-4 cores: Geochemistry, Geophysics, Geosystems, .08.006. straints on the Ginkgo Flow of the Columbia River v. 4, no. 2, p. 8705, doi:10.1029/2001GC000209. Kueppers, U., Perugini, D., and Dingwell, D.B., 2006b, Ex- Basalt Group: Geology, v. 25, p. 403–406, doi:10.1130 Kauahikaua, J., Mangan, M., Heliker, C., and Mattox, T., plosive energy during volcanic eruptions from fractal /0091-7613(1997)025<0403:TCOTGF>2.3.CO;2. 1996, A quantitative look at the demise of a basal- analysis of pyroclasts: Earth and Planetary Science Let- Hoblitt, R.P., 2000, Was the 18 May 1980 lateral blast at Mt. tic vent: The death of Kūpaianaha, Kilauea volcano, ters, v. 248, p. 800–807, doi:10.1016/j.epsl.2006.06.033. St. Helens the product of two explosions?: Philosophi- Hawai’i: Bulletin of Volcanology, v. 57, no. 8, p. 641– Lees, J.M., 1992, The magma system of Mount St. Helens: cal Transactions of the Royal Society of London, ser. A, 648, doi:10.1007/s004450050117. Non-linear high-resolution P-wave tomography: Jour- v. 358, p. 1639–1661, doi:10.1098/rsta.2000.0608. Kauahikaua, J., Cashman, K.V., Mattox, T.N., Heliker, C.C., nal of Volcanology and Geothermal Research, v. 53, Holness, M.B., and Humphreys, M.C.S., 2003, The Traigh Hon, K.A., Mangan, M.T., and Thornber, C.R., 1998, p. 103–116, doi:10.1016/0377-0273(92)90077-Q. Bhàn na Sgùrra Sill, Isle of Mull: Flow localization in Observations on basaltic lava streams in tubes from Lejeune, A.M., and Richet, P., 1995, Rheology of crystal- a major magma conduit: Journal of Petrology, v. 44, Kilauea volcano, Island of Hawai’i: Journal of Geo- bearing silicate melts: An experimental study at high p. 1961–1976, doi:10.1093/petrology/egg066. physical Research, v. 103, p. 27303–27323, doi:10.1029 viscosities: Journal of Geophysical Research, v. 100, Hon, K., Kauahikaua, J., Denlinger, R., and Mackay, K., /97JB03576. p. 4215–4229, doi:10.1029/94JB02985. 1994, Emplacement and infl ation of pahoehoe sheet Kauahikaua, J., Cashman, K.V., Clague, D.A., and Cham- Lesti, C., Porreca, M., Giordano, G., Mattei, M., Cas, R., Wright, fl ows: Observations and measurements of active lava pion, D., 2002, Geology of the most recent eruptions H., Folkes, C., and Viramonte, J., 2011, High-temperature fl ows on Kilauea volcano, Hawaii: Geological Society of Hualalai volcano, Hawaiì: Bulletin of Volcanology, emplacement of the Cerro Galán and Toconquis Group of America Bulletin, v. 106, p. 351–370, doi:10.1130 v. 64, p. 229–253, doi:10.1007/s00445-001-0196-8. ignimbrites (Puna Plateau, NW Argentina) determined /0016-7606(1994)106<0351:EAIOPS>2.3.CO;2. Kauahikaua, J., Sherrod, D.R., Cashman, K.V., Heliker, by TRM analyses: Bulletin of Volcanology, v. 73, no. 10, Hon, K., Gansecki, C., and Kauahikaua, J.P., 2003, The tran- C., Hon, K., Mattox, T.N., and Johnson, J.A., 2003, p. 1535–1565, doi:10.1007/s00445-011-0536-2. sition from ‘a‘ā to pāhoehoe crust on fl ows emplaced Hawaiian Lava-fl ow dynamics during the Pu‘u ‘Ō‘ō- Linde, A.T., Agustsson, K., Sacks, I.S., and Stefansson, S., during the Pu‘u ‘Ō‘ō-Kūpaianaha eruption, in Heliker, Kūpaianaha eruption: A tale of two decades, in Heliker, 1993, Mechanism of the 1991 eruption of Hekla from C.C., Swanson, D.A., and Takahashi, T.J., eds., The C.C., Swanson, D.A., and Takahashi, T.J., eds., 2003, continuous borehole strain monitoring: Nature, v. 365, Pu‘u ‘Ō‘ō-Kūpaianaha Eruption of Kīlauea Volcano, The Pu‘u ‘Ō‘ō-Kūpaianaha Eruption of Kīlauea Vol- p. 737–740, doi:10.1038/365737a0.

686 Geological Society of America Bulletin, May/June 2013 How volcanoes work

Lipman, P.W., 1997, Subsidence of ash-fl ow calderas: Re- Mastin, L.G., 2007, A user-friendly one-dimensional Murphy, M.D., Sparks, R.S.J., Barclay, J., Carroll, M.R., and lation to caldera size and magma-chamber geometry: model for wet volcanic plumes: Geochemistry, Geo- Brewer, T.S., 2000, Remobilization of andesite magma Bulletin of Volcanology, v. 59, p. 198–218, doi:10.1007 physics, Geosystems, v. 8, no. 3, Q03014, doi:10.1029 by intrusion of mafi c magma at the Soufriere Hills vol- /s004450050186. /2006GC001455. cano, Montserrat, West Indies: Journal of Petrology, Lipman, P.W., and Mullineaux, D.R., 1981, The 1980 Erup- Mastin, L.G., and Pollard, D.D., 1988, Surface deforma- v. 41, p. 21–42, doi:10.1093/petrology/41.1.21. tions of Mount St. Helens: U.S. Geological Survey tion and shallow dyke intrusion processes at Inyo Nakada, S., and Motomura, Y., 1999, Petrology of the Professional Paper 1250, 850 p. Craters, Long Valley, California: Journal of Geophysi- 1991–1995 eruption at Unzen: Effusion pulsation Lister, J.R., and Kerr, R.C., 1991, Fluid-mechanical models of cal Research, v. 93, p. 13,221–13,235, doi:10.1029 and groundmass crystallization: Journal of Volcanol- crack propagation and their application to magma trans- /JB093iB11p13221. ogy and Geothermal Research, v. 89, p. 173–196, doi: port in dykes: Journal of Geophysical Research, v. 96, Mastin, L.G., and 16 others, 2009a, A multidisciplinary effort 10.1016/S0377-0273(98)00131-0. no. B6, p. 10,049–10,077, doi:10.1029/91JB00600. to assign realistic source parameters to models of vol- Nakada, S., Shimizu, H., and Ohta, K., 1999, Overview of the Liu, Y., Anderson, A.T., Wilson, C.J.N., Davis, A.M., and canic ash-cloud transport and dispersion during erup- 1990–1995 eruption at Unzen Volcano: Journal of Vol- Steele, I.M., 2006, Mixing and differentiation in the tions: Journal of Volcanology and Geothermal Research, canology and Geothermal Research, v. 89, p. 1–22, doi: Oruanui rhyolitic magma, Taupo, New Zealand: Evi- v. 186, p. 10–21, doi:10.1016/j.jvolgeores.2009.01.008. 10.1016/S0377-0273(98)00118-8. dence from volatiles and trace elements in melt inclu- Mastin, L.G., Spieler, O., and Downey, W.S., 2009b, An Neri, A., Esposti Ongaro, T., Macedonio, G., and Gidaspow, sions: Contributions to Mineralogy and Petrology, experimental study of hydromagmatic fragmentation D., 2003, Multiparticle simulation of collapsing vol- v. 151, p. 71–87, doi:10.1007/s00410-005-0046-3. through energetic, non-explosive magma–water mix- canic columns and pyroclastic fl ow: Journal of Geo- Llewellin, E., and Manga, M., 2005, Bubble suspension ing: Journal of Volcanology and Geothermal Research, physical Research, v. 108, no. B4, p. 2202, doi:10.1029 rheology and implications for conduit fl ow: Journal of v. 180, p. 161–170. /2001JB000508. Volcanology and Geothermal Research, v. 143, p. 205– Melnik, O., 2000, Dynamics of two-phase conduit fl ow of Neuberg, J.W., Tuffen, H., Collier, H., Green, D., Powell, T., 217, doi:10.1016/j.jvolgeores.2004.09.018. high-viscosity gas-saturated magma: Large variations and Dingwell, D.B., 2006, The trigger mechanism of Lu, Z., Fielding, E., Patrick, M.R., and Trautwein, C.M., of sustained explosive eruption intensity: Bulletin of low-frequency earthquakes on Montserrat: Journal of 2003, Estimating lava volume by precision combina- Volcanology, v. 62, no. 3, p. 153–170, doi:10.1007/ Volcanology and Geothermal Research, v. 153, p. 37– tion of multiple baseline spaceborne and airborne inter- s004450000072. 50, doi:10.1016/j.jvolgeores.2005.08.008. ferometric synthetic aperture radar: The 1997 eruption Melnik, O., and Sparks, R.S.J., 1999, Nonlinear dynamics Newhall, C.G., and Dzurisin, D., 1989, Historical unrest at of Okmok volcano, Alaska: Geoscience and Remote of lava extrusion: Nature, v. 402, p. 37–41, doi:10.1038 large calderas of the world: Journal of Geology, v. 97, Sensing: IEEE Transactions, v. 41, p. 1428–1436. /46950. 650 p. Luhr, J., 2001, Glass inclusions and melt volatile contents Melnik, O., and Sparks, R.S.J., 2002, Dynamics of magma Newhall, C.G., and Punongbayan, R.S., eds., 1996, Fire at Parícutin volcano, Mexico: Contributions to Miner- ascent and lava extrusion at Soufriere Hills volcano, and Mud—Eruptions and Lahars of Mount Pinatubo, alogy and Petrology, v. 142, p. 261–283, doi:10.1007 Montserrat, in Druitt, T.H., and Kokelaar, B.P., eds., Philippines: Seattle, University of Washington Press, /s004100100293. The Eruption of Soufriere Hills Volcano, Montserrat, 1126 p. Luhr, J.F., and Simkin, T., 1993, Paricutín: The Volcano from 1995 to 1999: Geological Society of London Newhall, C.G., and Self, S., 1982, The volcanic explosivity Born in a Mexican Cornfi eld: Phoenix, Arizona, Geo- Memoir 21, p. 153–171. index (VEI) An estimate of explosive magnitude for science Press, Inc., 427 p. Melnik, O., and Sparks, R.S.J., 2005, Controls on con- historical volcanism: Journal of Geophysical Research, Lupi, M., Geiger, S., Carey, R.J., Thordarson, T., and Hough- duit magma fl ow dynamics lava-extrusion building v. 87, p. 1231–1238, doi:10.1029/JC087iC02p01231. ton, B.F., 2011, A model for syn-eruptive ground water eruptions: Journal of Geophysical Research, v. 110, Newman, S., and Lowenstern, J.B., 2002, VolatileCalc: A

flow during the phreatoplinian phase of the 28–29 B02209, doi:10.1029/2004JB003183. silicate melt–H2O–CO2 solution model written in Vis- March 1875 Askja volcano eruption, Iceland: Journal Melnik, O., Sparks, R.S.J., Costa, A., and Barmin, A., 2008, ual Basic for excel: Computers & Geosciences, v. 28, of Volcanology and Geothermal Research, v. 203, Volcanic eruptions: Cyclicity during lava dome growth, p. 597–604, doi:10.1016/S0098-3004(01)00081-4. p. 146–157, doi:10.1016/j.jvolgeores.2011.04.009. in Meyers, R.A., ed., Encyclopedia of Complexity and Nooner, S.L., and Chadwick, W.W., Jr., 2009, Volcanic in- Lyman, A.W., Koenig, E., and Fink, J.H., 2004, Predicting Systems Science: Springer, p. 1–22. fl ation measured in the caldera of Axial Seamount: yield strengths and effusion rates of lava domes from Menand, T., 2011, Physical controls and depth of emplace- Implications for magma supply and future eruptions: morphology and underlying topography: Journal of ment of igneous bodies: A review: Tectonophysics, Geochemistry, Geophysics, Geosystems, v. 10, no. 2, Volcanology and Geothermal Research, v. 129, p. 125– v. 500, p. 11–19, doi:10.1016/j.tecto.2009.10.016. Q02002, doi:10.1029/2008GC002315. 138, doi:10.1016/S0377-0273(03)00236-1. Métrich, N., and Wallace, P.J., 2008, Volatile abundances in Ofeigsson, B.G., Hooper, A., Sigmundsson, F., Sturkell, Mader, H.M., Zhang, Y., Phillips, J.C., Sparks, R.S.J., Stur- basaltic magmas and their degassing paths tracked by E., and Grapenthin, R., 2011, Deep magma storage at tevant, B., and Stolper, E., 1994, Experimental simula- melt inclusions: Reviews in Mineralogy and Geochem- Hekla volcano, Iceland, revealed by InSAR time series tions of explosive degassing of magma: Nature, v. 372, istry, v. 69, p. 363–402, doi:10.2138/rmg.2008.69.10. analysis: Journal of Geophysical Research, v. 116, p. 85–88, doi:10.1038/372085a0. Michaut, M., Bercovici, D., and Sparks, R.S.J., 2009, As- no. B5, B05401, doi:10.1029/2010JB007576. Manga, M., 1996, Waves of bubbles in basaltic magmas cent and compaction of gas rich magma and the ef- Ogden, D., Glatzmeier, G., and Wohletz, K., 2008, Effects and lavas: Journal of Geophysical Research, v. 101, fects of hysteretic permeability: Earth and Planetary of vent overpressure on buoyant eruption columns; p. 17,457–17,465, doi:10.1029/96JB01504. Science Letters, v. 282, p. 258–267, doi:10.1016/j.epsl implications for plume stability: Earth and Planetary Mangan, M., and Sisson, T., 2000, Delayed, disequilibrium .2009.03.026. Science Letters, v. 268, p. 283–292, doi:10.1016 degassing in rhyolite magma: Decompression experi- Mitchell, N., 2012, Submarine volcanism: Hot, cracking rocks /j.epsl.2008.01.014. ments and implications for explosive volcanism: Earth deep down: Nature Geoscience, v. 5, p. 444–445, doi: Okumura, S., and Nakamura, M., et al., 2009, Magma and Planetary Science Letters, v. 183, p. 441–455, 10.1038/ngeo1505. defor mation may induce non-explosive volcanism via doi:10.1016/S0012-821X(00)00299-5. Mora, J., Macias, J., Saucedo, R., Orlando, A., Manetti, degassing through bubble networks: Earth and Plan- Mangan, M., and Sisson, T., 2005, Evolution of melt-vapor P., and Vaselli, O., 2002, Petrology of the 1998–2000 etary Science Letters, v. 281, p. 267–274, doi:10.1016 surface tension in silicic volcanic systems: Experi- products of Volcán de Colima, México: Journal of Vol- /j.epsl.2009.02.036. ments with hydrous melts: Journal of Geophysical Re- canology and Geothermal Research, v. 117, p. 195– Ongaro, T.E., Widiwijayanti, C., Clarke, A.B., Voight, B., search, v. 110, B01202, doi:10.1029/2004JB003215. 212, doi:10.1016/S0377-0273(02)00244-5. and Neri, A., 2011, Multiphase-fl ow numerical model- Mangan, M., Mastin, L., and Sisson, T., 2004, Gas evolu- Moran, S., Malone, S., Qamar, A.I., Thelen, W., Wright, A.K., ing of the 18 May 1980 lateral blast at Mount St. Helens, tion in eruptive conduits: Combining insights from and Caplan-Auerbach, J., 2008, Seismicity associated USA: Geology, v. 39, no. 6, p. 535–538, doi:10.1130 high temperature and pressure decompression experi- with the renewed dome-building eruption of Mount St. /G31865.1. ments with steady-state fl ow modeling: Journal of Vol- Helens, 2004–2005, in Sherrod, D.R., Scott, W.E., and Oppenheimer, C., 1998, Volcanological applications canology and Geothermal Research, v. 129, p. 23–36, Stauffer, P.H., eds., A Volcano Rekindled: The Renewed of meteorological satellites: International Journal of doi:10.1016/S0377-0273(03)00230-0. Eruption of Mount St. Helens, 2004–2006: U.S. Geo- Remote Sensing, v. 19, p. 2829–2864, doi:10.1080 Mangan, M.T., and Cashman, K.V., 1996, The structure of logical Survey Professional Paper 1750, p. 27–60. /014311698214307. basaltic scoria and reticulite and inferences for vesicu- Moran, S., Newhall, C., and Roman, D., 2011, Failed Oppenheimer, C., 2003, Climatic, environmental and hu- lation, foam formation, and fragmentation in lava foun- magmatic eruptions: Late-stage cessation of magma man consequences of the largest known historic erup- tains: Journal of Volcanology and Geothermal Research, ascent: Bulletin of Volcanology, v. 73, p. 115–122, tion: Tambora volcano (Indonesia) 1815: Progress in v. 73, p. 1–18, doi:10.1016/0377-0273(96)00018-2. doi:10.1007/s00445-010-0444-x. Physical Geography, v. 27, p. 230–259, doi:10.1191 Mason, B.G., Pyle, D.M., and Oppenheimer, C., 2004, Mudde, R.F., 2005, Gravity-driven bubbly fl ows: Annual Re- /0309133303pp379ra. The size and frequency of the largest explosive erup- view of Fluid Mechanics, v. 37, p. 393–423, doi:10.1146 Pal, R., 2003, Rheological behavior of bubble-bearing tions: Bulletin of Volcanology, v. 66, p. 735–748, /annurev.fl uid.37.061903.175803. magmas: Earth and Planetary Science Letters, v. 207, doi:10.1007/s00445-004-0355-9. Mueller, S., Scheu, B., Spieler, O., and Dingwell, D., 2008, p. 165–179, doi:10.1016/S0012-821X(02)01104-4. Mason, R.M., Starostin, A.B., Melnik, O.E., and Sparks, Permeability control on magma fragmentation: Geol- Pallister, J.S., Thornber, C.R., Cashman, K.V., Clynne, R.S.J., 2006, From Vulcanian explosions to sustained ogy, v. 36, p. 399–402, doi:10.1130/G24605A.1. M.A., Lowers, H.A., Brownfi eld, I.K., and Meeker, explosive eruptions: The role of diffusive mass trans- Mueller, S., Llewellin, E.W., and Mader, H.M., 2010, The G.P., 2008, Petrology of the 2004–2006 Mount St. fer in conduit fl ow dynamics: Journal of Volcanol- rheology of suspensions of solid particles: Proceedings Helens lava dome—Implications for magmatic plumb- ogy and Geothermal Research, v. 153, p. 148–165, of the Royal Society A: Mathematical, Physical and ing, explosivity and eruption triggering, in Sherrod, doi:10.1016/j.jvolgeores.2005.08.011. Engineering Science, v. 466, p. 1201–1228. D.R., Scott, W.E., and Stauffer, P.H., eds., A Volcano

Geological Society of America Bulletin, May/June 2013 687 Cashman and Sparks

Rekindled: The Renewed Eruption of Mount St. Hel- Oregon, revealed by InSAR: Geochemistry, Geo- strength in crystal-melt suspensions: Earth and Plan- ens, 2004–2006: U.S. Geological Survey Professional physics, Geosystems, v. 12, Q12005, doi:10.1029 etary Science Letters, v. 187, p. 367–379, doi:10.1016 Paper 1750, p. 647–702. /2011GC003826. /S0012-821X(01)00289-8. Papale, P., Moretti, R., and Barbato, D., 2006, The compo- Riehle, J.R., Miller, T.F., and Bailey, R.A., 1995, Cooling, Sable, J.E., Houghton, B.F., Del Carlo, P., and Coltelli,

sitional dependence of the saturation surface of H2O- degassing and compaction of rhyolitic ash fl ow tuffs: A M., 2006, Changing conditions of magma ascent and CO2 fl uids in silicate melts: Chemical Geology, v. 229, computational model: Bulletin of Volcanology, v. 57, fragmentation during the Etna 122 BC basaltic Plinian p. 78–95, doi:10.1016/j.chemgeo.2006.01.013. no. 5, p. 319–336. eruption: Evidence from clast microtextures: Journal of Paton, D., and Johnston, D.M., 2006, Disaster Resilience: Riker, J.M., Cashman, K.V., and Kauahikaua, J.P., 2009, Volcanology and Geothermal Research, v. 158, p. 333– An Integrated Approach: Springfi eld, Illinois, Charles The length of channelized lava fl ows: Insight from the 354, doi:10.1016/j.jvolgeores.2006.07.006. C. Thomas, 320 p. 1859 eruption of Mauna Loa volcano, Hawaii: Jour- Sahimi, M., 1994, Applications of Percolation Theory: Lon- Patra, A.K., Bauer, A.C., Nichita, C.C., Pitman, E.B., Sheri- nal of Volcanology and Geothermal Research, v. 183, don, Taylor & Francis, 276 p. dan, M.F., Bursik, M., Rupp, B., Webber, A., Stinton, p. 139–156, doi:10.1016/j.jvolgeores.2009.03.002. Sakuma, S., Kajiwara, T., Nakada, S., Uto, K., and Shimizu, A.J., Namikawa, L.M., and Renschler, C.S., 2005, Parallel Ripepe, M., Marchetti, E., Ulivieri, G., Harris, A., Dehn, H., 2008, Drilling and logging results of USDP-4— adaptive numerical simulation of dry avalanches over J., Burton, M., Caltabiano, T., and Salerno, G., 2005, Penetration into the volcanic conduit of Unzen volcano, natural terrain: Journal of Volcanology and Geothermal Effusive to explosive transition during the 2003 erup- Japan : Journal of Volcanology and Geothermal Research, Research, v. 139, p. 1–21, doi:10.1016/j.jvolgeores tion of Stromboli volcano: Geology, v. 33, p. 341–344, v. 175, no. 1–2, p. 1–12, doi:10.1016/j.jvolgeores.2008 .2004.06.014. doi:10.1130/G21173.1. .03.039. Paulatto, M., Annen, C., Henstock, T.J., Kiddle, E., Minshull , Roche, O., Montserrat, S., Nio, Y., and Tamburrino, A., Santacroce, R., Bertagnini, A., Civetta, L., Landi, P., and T.A., Sparks, R.S.J., and Voight, B., 2012, Magma 2010, Pore fl uid pressure and internal kinematics of Sbrana, A., 1993, Eruptive dynamics and petrogenetic chamber properties from integrated seismic tomogra- gravitational laboratory air-particle fl ows: Insights processes in a very shallow magma reservoir: The phy and thermal modeling at Montserrat: Geochem- into the emplacement dynamics of pyroclastic fl ows: 1906 eruption of Vesuvius: Journal of Petrology, v. 34, istry, Geophysics, Geosystems, v. 13, Q01014, doi: Journal of Geophysical Research, v. 115, B09206, p. 383–425. 10.1029/2011GC003892. doi:10.1029/2009JB007133. Saunders, K., Blundy, J., Dohmen, R., and Cashman, K., Pearse, J., and Fialko, Y., 2010, Mechanics of active mag- Roggensack, K., Hervig, R.L., McKnight, S.B., and Wil- 2012, Linking petrology and seismology at an active matic intraplating in the Rio Grande Rift near Socorro, liams, S.N., 1997, Explosive basaltic volcanism from volcano: Science, v. 336, p. 1023–1027, doi:10.1126 New Mexico: Journal of Geophysical Research, v. 115, Cerro Negro volcano: Infl uence of volatiles on erup- /science.1220066. B07413, doi:10.1029/2009JB006592. tive style: Science, v. 277, p. 1639–1642, doi:10.1126 Scandone, R., and Malone, S.D., 1985, Magma supply, Pérez, W., and Freundt, A., 2006, The youngest highly ex- /science.277.5332.1639. magma discharge and readjustment of the feeding sys- plosive basaltic eruptions from Masaya caldera (Nica- Roman, D.C., and Cashman, K.V., 2006, The origin of vol- tem of Mount St. Helens during 1980: Journal of Vol- ragua), in Rose, W.I., Bluth, G.J.S., Carr, M.J., Ewert, cano tectonic earthquake swarms: Geology, v. 34, p. 457– canology and Geothermal Research, v. 23, p. 239–262, J.W., Patino, L.C., and Vallance, J.W., eds., Volcanic 460, doi:10.1130/G22269.1. doi:10.1016/0377-0273(85)90036-8. Hazards in Central America: Geological Society of Roman, D.C., Neuberg, J., and Luckett, R.R., 2006, As- Scandone, R., Cashman, K.V., and Malone, S.D., 2007, America Special Paper 412, p. 189–207. sessing the likelihood of volcanic eruption through Magma supply, magma ascent and the style of volcanic Peterson, D.W., Holcomb, R.T., Tilling, R.I., and Chris- analysis of volcanotectonic earthquake fault-plane eruptions: Earth and Planetary Science Letters, v. 253, tiansen, R.L., 1994, Development of lava tubes in the solutions: Earth and Planetary Science Letters, v. 248, p. 513–529, doi:10.1016/j.epsl.2006.11.016. light of observations at Mauna Ulu, Kilauea volcano, p. 244–252, doi:10.1016/j.epsl.2006.05.029. Scandone, R., Barberi, F., and Rosi, M., 2009, The 2007 Hawaii : Bulletin of Volcanology, v. 56, p. 343–360. Rowland, S.K., and Walker, G.P.L., 1990, Pahoehoe and eruption of Stromboli: Preface: Journal of Volcanology Pichavant, M., Costa, F., Burgisser, A., Scaillet, B., Martel, aa in Hawaii: Volumetric fl ow rate controls the lava and Geothermal Research, v. 182, p. 3–4, doi:10.1016 C., and Poussineau, S., 2007, Equilibration scales in structure: Bulletin of Volcanology, v. 52, p. 631–641, /j.jvolgeores.2008.12.019. silicic to intermediate magmas—Implications for ex- doi:10.1007/BF00301212. Self, S., 1992, Krakatau revisited: The course of events and perimental studies: Journal of Petrology, v. 48, p. 1955– Rubin, K.H., Soule, S.A., Chadwick, W.W.J., Fornari, D.J., interpretation of the 1883 eruption: GeoJournal, v. 28, 1972, doi:10.1093/petrology/egm045. Clague, D.A., Embley, R.W., Baker, E.T., Perfi t, M.R., p. 109–121, doi:10.1007/BF00177223. Pinel, V., and Jaupart, C., 2004, Magma storage and hori- Caress, D.W., and Dziak, R.P., 2012, Volcanic erup- Sherrod, D.R., Scott, W.E., and Stauffer, P.H., eds., 2008, A zontal dyke injection beneath a volcanic edifi ce: Earth tions in the deep sea: Oceanography, v. 25, p. 142–157. Volcano Rekindled; the Renewed Eruption of Mount and Planetary Science Letters, v. 221, p. 245–262, Ruprecht, P., and Cooper, K.M., 2012, Integrating the ura- St. Helens, 2004–2006: U.S. Geological Survey Pro- doi:10.1016/S0012-821X(04)00076-7. nium-series and elemental diffusion geochronometers fessional Paper 1750, 856 p. and DVD-ROM, http:// Pinkerton, H., and Wilson, L., 1994, Factors controlling the in mixed magmas from Volcán Quizapu, Central Chile: pubs.usgs.gov/pp/1750/. lengths of channel-fed lava fl ows: Bulletin of Volcanol- Journal of Petrology, v. 53, p. 801–840, doi:10.1093 Sigmundsson, F., and 15 others, 2010, Intrusion triggering ogy, v. 56, p. 108–120. /petrology/egs001. of the 2010 Eyjafjallajökull explosive eruption: Nature, Pioli, L., Erlund, E., Johnson, E., Cashman, K.V., Wallace, Rust, A.C., and Cashman, K.V., 2004, Permeability and v. 468, p. 426–430, doi:10.1038/nature09558. P., Rosi, M., and Delgado Granados, H., 2008, Explo- degassing of vesicular silicic magma: Earth and Plan- Sigurdsson, H., Houghton, B., Rymer, H., and Stix, J., eds., sive dynamics of violent Strombolian eruptions: The etary Science Letters, v. 228, p. 93–107, doi:10.1016/j 2000, Encyclopedia of Volcanoes: New York, Aca- eruption of Parícutin volcano 1943–1952 (Mexico): .epsl.2004.09.025. demic Press, 1389 p. Earth and Planetary Science Letters, v. 271, p. 359– Rust, A.C., and Cashman, K.V., 2007, Multiple origins of Simkin, T., and Fiske, R.S., 1983, Krakatau 1883: The 368, doi:10.1016/j.epsl.2008.04.026. obsidian pyroclasts and implications for changes in the Volcanic Eruption and its Effects: Washington, D.C., Pioli, L., Azzopardi, B.J., and Cashman, K.V., 2009, Con- dynamics of the 1300 B.P. eruption of Newberry vol- Smithsonian Institution Press. trols on the explosivity of scoria cone eruptions: cano, USA: Bulletin of Volcanology, v. 69, p. 825–845, Singh, S.C., Harding, A.J., Kent, G.M., Sinha, M.C., Combier, Magma segregation at conduit junctions: Journal of doi:10.1007/s00445-006-0111-4. V., Bazin, S., Tong, C.H., Pye, J.W., Barton, P.J., Hobbs, Volcanology and Geothermal Research, v. 186, p. 407– Rust, A.C., and Cashman, K.V., 2011, Permeability con- R.W., White, R.S., and Orcutt, J.A., 2006, Seismic refl ec- 415, doi:10.1016/j.jvolgeores.2009.07.014. trols on expansion and size distributions of pyroclasts: tion images of the Moho underlying melt sills at the East Pioli, L., Bonadonna, C., Azzopardi, B.J., Phillips, J.C., and Journal of Geophysical Research, v. 116, B11202, Pacifi c Rise: Nature, v. 442, p. 287–290, doi:10.1038 Ripepe, M., 2012, Experimental constraints on the doi:10.1029/2011JB008494. /nature04939. outgassing dynamics of basaltic magmas: Journal of Rust, A.C., and Manga, M., 2002, Bubble shapes and orien- Slack, P.D., Fox, C.G., and Dziak, R.P., 1999, P wave detec- Geophysical Research, v. 117, B03204, doi:10.1029 tations in low Re simple shear fl ow: Journal of Colloid tion thresholds, Pn velocity estimates, and T wave lo- /2011JB008392. and Interface Science, v. 249, p. 476–480, doi:10.1006 cation uncertainty from oceanic hydrophones: Journal Pritchard, M.E., and Simons, M., 2004, An InSAR-based /jcis.2002.8292. of Geophysical Research, v. 104, no. B6, p. 13,061– survey of volcanic deformation in the central Andes: Rust, A.C., Manga, M., and Cashman, K.V., 2003, Deter- 13,072, doi:10.1029/1999JB900112. Geochemistry, Geophysics, Geosystems, v. 5, no. 2, mining fl ow type, shear rate and shear stress in magmas Slezin, Y.B., 2003, The mechanism of volcanic eruption Q02002, doi:10.1029/2003GC000610. from bubble shapes and orientations: Journal of Vol- (steady state approach): Journal of Volcanology and Putirka, K.D., 2008, Thermometers and barometers for volcanology and Geothermal Research, v. 122, no. 1–2, Geothermal Research, v. 122, p. 7–50, doi:10.1016 canic systems: Reviews in Mineralogy and Geochemis- p. 111–132, doi:10.1016/S0377-0273(02)00487-0. /S0377-0273(02)00464-X. try, v. 69, no. 1, p. 61–120, doi:10.2138/rmg.2008.69.3. Rust, A.C., Cashman, K.V., and Wallace, P.J., 2004, Magma Smith, V.C., Blundy, J.D., and Arce, J.L., 2009, A temporal Pyle, D.M., 1989, The thickness, volume and grainsize of degassing buffered by vapor fl ow through brecciated record of magma accumulation and evolution beneath tephra fall deposits: Bulletin of Volcanology, v. 51, conduit margins: Geology, v. 32, p. 349–352, doi: Nevado de Toluca, Mexico, preserved in plagioclase p. 1–15, doi:10.1007/BF01086757. 10.1130/G20388.2 phenocrysts: Journal of Petrology, v. 50, p. 405–426, Realmuto, V.J., Hon, K., Kahle, A.B., Abbott, E.A., and Ryan, M., 1987, Neutral buoyancy and the mechanical evo- doi:10.1093/petrology/egp005. Pieri, D.C., 1992, Multispectral thermal infrared map- lution of magmatic systems, in Mysen, B.O., ed., Mag- Soule, S.A., and Cashman, K.V., 2005, The shear rate de- ping of the 1 October 1988 Kupaianaha fl ow fi eld, matic Processes: Physiochemical Principles, A Volume pendence of the pahoehoe-to-a’a transition: Analog Kilauea volcano, Hawaii: Bulletin of Volcanology, in Honor of Hatten S. Yoder, Jr.: The Geochemical So- experiments: Geology, v. 33, p. 361–364, doi:10.1130 v. 55, p. 33–44, doi:10.1007/BF00301118. ciety, p. 259–288. /G21269.1. Riddick, S.N., and Schmidt, D.A., 2011, Time-dependent Saar, M.O., Manga, M., Cashman, K.V., and Fremouw, Soule, S.A., Cashman, K.V., and Kauahikaua, J.P., 2004, changes in volcanic infl ation rate near Three Sisters, S., 2001, Numerical models of the onset of yield Examining fl ow emplacement through the surface

688 Geological Society of America Bulletin, May/June 2013 How volcanoes work

morphology of three rapidly emplaced, solidifi ed lava Taddeucci, J., Pompilio, M., and Scarlato, P., 2004, Conduit low magmatic and hydrothermal systems: Journal of fl ows, Kilauea volcano, Hawai’i: Bulletin of Volcanol- processes during the July–August 2001 explosive ac- Geophysical Research, v. 113, B02305, doi:10.1029 ogy, v. 66, p. 1–14, doi:10.1007/s00445-003-0291-0. tivity of Mt. Etna (Italy): Inferences from glass chemis- /2007JB005259. Spampinato, L., Calvari, S., Oppenheimer, C., and Boschi, try and crystal size distribution of ash particles: Journal Walker, G.P.L., Huntingdon, A.T., Sanders, A.T., and Dins- E., 2011, Volcano surveillance using infrared cameras: of Volcanology and Geothermal Research, v. 137, dale, J.L., 1973, Lengths of lava fl ows: Philosophical Earth-Science Reviews, v. 106, p. 63–91. p. 33–54, doi:10.1016/j.jvolgeores.2004.05.011. Transactions of the Royal Society of London, ser. A, Sparks, R.S.J., 1978, The dynamics of bubble formation and Taisne, B., and Jaupart, C., 2009, Dike propagation through Mathematical and Physical Sciences, v. 274, no. 1238, growth in magmas: A review and analysis: Journal of layered rocks: Journal of Geophysical Research, p. 107–118, doi:10.1098/rsta.1973.0030. Volcanology and Geothermal Research, v. 3, p. 1–37, v. 114, B09203, doi:10.1029/2008JB006228. Wallace, P.J., Anderson, A.T., Jr., and Davis, A.M., 1999,

doi:10.1016/0377-0273(78)90002-1. Taisne, B., and Tait, S., 2009, Eruption versus intrusion? Gradients in H2O, CO2, and exsolved gas in a large- Sparks, R.S.J., 1997, Causes and consequences of pressuri- Arrest of propagation of constant volume, buoyant, volume silicic magma system: Interpreting the record sation in lava dome eruptions: Earth and Planetary Sci- liquid-fi lled cracks in an elastic, brittle host: Journal preserved in melt inclusions from the Bishop Tuff: ence Letters, v. 150, p. 177–189, doi:10.1016/S0012 of Geophysical Research, v. 114, B06202, doi:10.1029 Journal of Geophysical Research, v. 104, p. 20097– -821X(97)00109-X. /2009JB006297. 20122, doi:10.1029/1999JB900207. Sparks, R.S.J., and Young, S.R., 2002, The eruption of Sou- Taisne, B., and Tait, S., 2011, Effect of solidifi cation on a Walsh, S.D.C., and Saar, M.O., 2008, Numerical models of frière Hills volcano, Montserrat: Overview of scientifi c propagating dike: Journal of Geophysical Research, stiffness and yield stress growth in crystal-melt sus- results, in Druitt, T.H., and Kokelaar, B.P., eds., The v. 116, B01206, doi:10.1029/2009JB007058. pensions: Earth and Planetary Science Letters, v. 267, Eruption of Soufrière Hills Volcano, Montserrat, from Tait, S., Jaupart, C., and Vergniolle, S., 1989, Pressure, gas con- p. 32–44, doi:10.1016/j.epsl.2007.11.028. 1995 to 1999: Geological Society of London Memoir tent and eruption periodicity of a shallow crystallizing Watts, R.B., Herd, R.A., Sparks, R.S.J., and Young, S.R., 21, p. 45–69. magma chamber: Earth and Planetary Science Letters, 2002, Growth patterns and emplacement of the ande- Sparks, R.S.J., Bursik, M.I., Carey, S.N., Gilbert, J.S., v. 92, p. 107–123, doi:10.1016/0012-821X(89)90025-3. site lava dome at the Soufriere Hills volcano, Mont- Glaze, L., Sigurdsson, H., and Woods, A.W., 1997, Tani, K., Fiske, R., Tamura, Y., Kido, Y., Naka, J., Shu- serrat, in Druitt, T.H., and Kokelaar, B.P., eds., The Volcanic Plumes: John Wiley and Sons, 557 p. kuno, H., and Takeuchi, R., 2008, Sumisu volcano, Eruption of the Soufrière Hills Volcano, Montserrat Sparks, R.S.J., Tait, S.R., and Yanev, Y., 1999, Dense Izu-Bonin arc, Japan: Site of a silicic caldera-forming 1995 to 1999: Geological Society of London Mem- welding caused by volatile resorption: Journal of the eruption from a small open-ocean island: Bulletin of oir 21, p. 115–152. Geological Society of London, v. 156, p. 217–225, Volcanology, v. 70, p. 547–562, doi:10.1007/s00445 Webb, S.L., and Dingwell, D.B., 1990, The onset of non- doi:10.1144/gsjgs.156.2.0217. -007-0153-2. Newtonian rheology of silicate melts: Physics and Sparks, R.S.J., Brown, R.J., Field, M., and Gilbertson, M., Tarasewicz, J., Brandsdottir, B., White, R.S., Hensch, M., Chemistry of Minerals, v. 17, p. 125–132, doi:10.1007 2007, Kimberlite ascent and eruption: Nature, v. 450, and Thorbjarnardottir, B., 2012, Using microearth- /BF00199663. p. E21–E21, doi:10.1038/nature06435. quakes to track repeated magma intrusions beneath White, J.D.L., 2000, Subaqueous eruption-fed density cur- Sparks, R.S.J., Aspinall, W.P., Crosweller, H.S. and Hincks, the Eyjafjallajökull stratovolcano, Iceland: Journal of rents and their deposits: Precambrian Research, v. 101, T.K., 2012, Risk and uncertainty assessment of volcanic Geophysical Research, v. 117, B00C06, doi:10.1029 p. 87–109, doi:10.1016/S0301-9268(99)00096-0. hazards, in Hill, L., Rougier, J., and Sparks, R.S.J., eds., /2011JB008751. White, J.D.L., and Ross, P.S., 2011, Maar-diatreme vol canoes: Assessment of Risk and Uncertainty for Natural Haz- Toramaru, A., Noguchia, S., Oyoshiharab, S., and Tsunea, A review: Journal of Volcanology and Geothermal ards: Cambridge, Cambridge University Press (in press). A., 2008, MND (microlite number density) water ex- Research, v. 201, p. 1–29, doi:10.1016/j.jvolgeores Spieler, O., Kennedy, B., Kueppers, U., Dingwell, D.B., solution rate meter: Journal of Volcanology and Geo- .2011.01.010. Scheu, B., and Taddeucci, J., 2004, The fragmenta- thermal Research, v. 175, no. 1–2, p. 156–167, doi: White, J.D.L., Smellie, J.L., and Clague, D.A., eds., 2003, tion threshold of pyroclastic rocks: Earth and Plan- 10.1016/j.jvolgeores.2008.03.035. Explosive Subaqueous Volcanism: Geophysical Mono- etary Science Letters, v. 226, no. 1–2, p. 139–148, Tuffen, H., Dingwell, D.B., and Pinkerton, H., 2003, Re- graph 140: Washington, D.C., American Geophysical doi:10.1016/j.epsl.2004.07.016. peated fracture and healing of silicic magma generate Union, 379 p. Stasiuk, M.V., Jaupart, C., and Sparks, R.S.J., 1993, Varia- fl ow banding and earthquakes?: Geology, v. 31, no. 12, White, S.M., Crisp, J.A., and Spera, F.J., 2006, Long- tions of fl ow rate and volume during eruption of lava: p. 1089–1092, doi:10.1130/G19777.1. term volumetric eruption rates and magma budgets: Earth and Planetary Science Letters, v. 114, p. 505– Vergniolle, S., and Jaupart, C., 1986, Separated two-phase Geochemistry, Geophysics, Geosystems, v. 7, no. 3, 516, doi:10.1016/0012-821X(93)90079-O. fl ow and basaltic eruptions: Journal of Geophysical Q03010, doi:10.1029/2005GC001002. Stelling, P.S., Beget, J.B., Nye, C.N., Gardner, J.G., Devine, Research, v. 91, p. 12,842–12,860. Williams, S.N., 1983, Plinian airfall deposits of basaltic J.D., and George, R.G., 2002, Geology and petrology Vinkler, A.P., Cashman, K., Giordano, G., and Groppelli, G., composition: Geology, v. 11, p. 211–214, doi:10.1130 of ejecta from the 1999 eruption of Shishaldin volcano, 2012, Evolution of the mafi c Villa Senni caldera-form- /0091-7613(1983)11<211:PADOBC>2.0.CO;2. Alaska: Bulletin of Volcanology, v. 64, p. 548–561, ing eruption at Colli Albani volcano, Italy, indicated Wilson, C., and Hildreth, W., 1997, The Bishop Tuff: New doi:10.1007/s00445-002-0229-y. by textural analysis of juvenile fragments: Journal of insights from eruptive stratigraphy: The Journal of Geol- Stevens, N.F., Murray, J.B., and Wadge, G., 1997, The vol- Volcanology and Geothermal Research, v. 235–236, ogy, v. 105, p. 407–440, doi:10.1086/515937. ume and shape of the 1991–1993 lava fl ow fi eld at p. 37–54, doi:10.1016/j.jvolgeores.2012.03.006. Wilson, C.J.N., 1985, The Taupo eruption, New Zealand: Mount Etna, Sicily: Bulletin of Volcanology, v. 58, Voight, B., and Sparks, R.S.J., 2010, Introduction to spe- II. The Taupo ignimbrite: Philosophical Transactions p. 449–454, doi:10.1007/s004450050153. cial section on the “Eruption of Soufriere Hills Vol- of the Royal Society of London, ser. A, v. 314, p. 229– Stevens, N.F., Wadge, G., and Williams, C.A., 2001, Post- cano, Montserrat, the CALIPSO Project, and the 310, doi:10.1098/rsta.1985.0020. emplacement lava subsidence and the accuracy of ERS SEA-CALIPSO Arc-Crust Imaging Experiment”: Wilson, C.J.N., and Charlier, B.L.A., 2009, Rapid rates InSAR digital elevation models of volcanoes: Inter- Geophysical Research Letters, v. 37, L00E23, doi: of magma generation at contemporaneous magma national Journal of Remote Sensing, v. 22, p. 819–828, 10.1029/2010GL044254. systems, Taupo volcano, New Zealand: Insights doi:10.1080/01431160051060246. Voight, B., and 24 others, 1999, Magma fl ow instability and from U-Th model-age spectra in zircons: Journal of Stolper, E.M., DePaolo, D.J., and Thomas, D.M., 2009, Deep cyclic activity at Soufriere Hills volcano, Montserrat, Petrology, v. 50, p. 875–907, doi:10.1093/petrology drilling into a mantle plume volcano: The Hawaii Sci- B.W.I.: Science, v. 283, p. 1138–1142, doi:10.1126 /egp023. entifi c Drilling Project: Scientifi c Drilling, v. 7, p. 4–14. /science.283.5405.1138. Woods, A.W., 1995, The dynamics of explosive volcanic Suckale, J., Hager, B.H., Elkins-Tanton, L.T., and Nave, Voight, B., Komorowski, J.-C., Norton, G.E., Belousov, A.B., eruptions: Reviews of Geophysics, v. 33, no. 4, p. 495– J.-C., 2010, It takes three to tango: 2. Bubble dynamics Belousova, M., Boudon, G., Francis, P.W., Franz, W., 530, doi:10.1029/95RG02096. in basaltic volcanoes and ramifi cations for modeling Heinrich, P., Sparks, R.S.J., and Young, S.R., 2002, The Wright, H.M.N., Roberts, J.J., and Cashman, K.V., 2006, normal Strombolian activity: Journal of Geophysical 26 December (Boxing Day) 1997 sector collapse and Permeability of anisotropic tube pumice: Model calcu- Research, v. 115, B07410, doi:10.1029/2009JB006917. debris avalanche at Soufrière Hills volcano, Montserrat, lations and measurements: Geophysical Research Let- Swanson, D.A., 2008, Hawaiian oral tradition describes in Druitt, T.H., and Kokellar, B.P., eds., The eruption of ters, v. 33, no. 17, doi:10.1029/2006GL027224. 400 years of volcanic activity at Kīlauea: Journal of Soufriere Hills Volcano, Montserrat from 1995 to 1999: Wright, H.M.N., Cashman, K.V., Rosi, M., and Cioni, R., Volcanology and Geothermal Research, v. 176, no. 3, Geological Society of London Memoir 21, p. 363–407. 2007, Breadcrust bombs as indicators of Vulcanian p. 427–431, doi:10.1016/j.jvolgeores.2008.01.033. Wadge, G., Macfarlane, D.G., Odbert, H.M., James, M.R., eruption dynamics at Guagua Pichincha volcano, Ecua- Swanson, D.A., and Holcomb, R.T., 1990, Regularities Hole, J.K., Ryan, G., and Bass, V., 2008, Time series dor: Bulletin of Volcanology, v. 69, no. 3, p. 281–300, in the growth of the Mount St. Helens dacite dome, radar observations of a growing lava dome: Jour- doi:10.1007/s00445-006-0073-6. 1980–1986, in Fink, J., eds., Lava Flows and Domes. nal of Geophysical Research, v. 113, B08210, doi: Wright, H.M.N., Folkes, C.B., Cas, R.A.F., and Cashman, International Association of Volcanology and Chemis- 10.1029/2007JB005466. K.V., 2011, Heterogeneous pumice populations in try of the Earth’s Interior Proceedings in Volcanology: Waite, G.P., and Moran, S.C., 2009, VP structure of Mount the 2.56 Ma Cerro Galan ignimbrite: Implications for Heidelberg, Germany, Springer-Verlag, p. 3–24. St. Helens, Washington, USA, imaged with local earth- magma recharge and ascent preceding a large volume Swanson, D.A., Rose, T.R., Fiske, R.S., and McGeehin, J.P., quake tomography: Journal of Volcanology and Geo- silicic eruption: Bulletin of Volcanology, doi:10.1007 2012, Keanakāko‘i Tephra produced by 300 years of thermal Research, v. 182, no. 1–2, p. 113–122, doi: /s00445-011-0525-5. explosive eruptions following collapse of Kīlauea’s 10.1016/j.jvolgeores.2009.02.009. Wright, H.M.N., Cashman, K.V., Mothes, P.A., Hall, M.L., caldera in about 1500 CE: Journal of Volcanology Waite, G.P., Chouet, B.A., and Dawson, P.B., 2008, Erup- Ruiz, A.G., and Le Pennec, J.-L., 2012, Estimat- and Geothermal Research, v. 215–216, p. 8–25, doi: tion dynamics at Mount St. Helens imaged from ing rates of decompression from textures of erupted 10.1016/j.jvolgeores.2011.11.009. broadband seismic waveforms: Interaction of the shal- ash particles produced by 1999–2006 eruptions of

Geological Society of America Bulletin, May/June 2013 689 Cashman and Sparks

Tungura hua volcano, Ecuador: Geology, v. 40, p. 619– Wylie, J.J., Voight, B., and Whitehead, J.A., 1999, Instabil- tion timescales beneath Montserrat: Insights from Sr 622, doi:10.1130/G32948.1. ity of magma fl ow from volatile-dependent viscosity: and Ba zonation in plagioclase phenocrysts: Journal Wright, I., and Gamble, J.A., 1999, Southern Kermadec Science, v. 285, p. 1883–1885, doi:10.1126/science of Petrology, v. 44, no. 8, p. 1413–1431, doi:10.1093 submarine caldera arc volcanoes (SW Pacifi c): Cal- .285.5435.1883. /petrology/44.8.1413. dera formation by effusive and pyroclastic eruption: Zandt, G., Leidig, M., Chmielowski, J., Baumont, D., and Zimanowski, B., Wohletz, K., Dellino, P., and Büttner, Marine Geology, v. 161, p. 207–227, doi:10.1016 Yuan, X., 2003, Seismic detection and characteriza- R., 2003, The volcanic ash problem: Journal of Vol- /S0025-3227(99)00040-7. tion of the Altiplano-Puna magma body, central Andes: canology and Geothermal Research, v. 122, p. 1–5, Wright, I., Gamble, J., and Shane, P., 2003, Submarine silicic Pure and Applied Geophysics, v. 160, p. 789–807, doi:10.1016/S0377-0273(02)00471-7 volcanism of the Healy caldera, southern Kermadec doi:10.1007/PL00012557. arc (SW Pacifi c): I—Volcanology and eruption mecha- Zebker, H.A., Rosen, P., Hensley, S., and Mouginis-Mark, SCIENCE EDITOR: BRENDAN MURPHY nisms: Bulletin of Volcanology, v. 65, no. 1, p. 15–29. P.J., 1996, Analysis of active lava fl ows on Kilauea Wright, R., Flynn, L.P., Garbeil, H., Harris, A.J.L., and Pilger, volcano, Hawaii, using SIR-C radar correlation mea- MANUSCRIPT RECEIVED 30 APRIL 2012 E., 2004, MODVOLC: Near-real-time thermal moni- surements: Geology, v. 24, p. 495–498, doi:10.1130 REVISED MANUSCRIPT RECEIVED 10 AUGUST 2012 toring of global volcanism: Journal of Volcanology and /0091-7613(1996)024<0495:AOALFO>2.3.CO;2. MANUSCRIPT ACCEPTED 27 AUGUST 2012 Geothermal Research, v. 135, p. 29–49, doi:10.1016 Zellmer, G.F., Sparks, R.S.J., Hawksworth, C.J., and Wieden- /j.jvolgeores.2003.12.008. beck, M., 2003, Magma emplacement and remobiliza- Printed in the USA

690 Geological Society of America Bulletin, May/June 2013