“Poseidic” Explosive Eruptions at Loihi Seamount, Hawaii

Downloaded from geology.gsapubs.org on October 5, 2010 “Poseidic” explosive eruptions at Loihi Seamount, Hawaii

C. Ian Schipper*1, James D.L. White1, Bruce F. Houghton2, Nobumichi Shimizu3, and Robert B. Stewart4 1Geology Department, University of Otago, PO Box 56, Leith Street, Dunedin 9016, New Zealand 2School of Ocean and Earth Science and Technology (SOEST), University of Hawai’ i at Ma¯noa, 1680 East-West Road, Honolulu, Hawaii 98622, USA 3Geology and Geophysics, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543, USA 4Soil and Earth Sciences, Institute of Natural Resources (INR), Massey University, PB 11-222, Palmerston North 4474, New Zealand

ABSTRACT (A.D. 1996) of Loihi’s ~400 ka history (Moore Much remains unknown about submarine explosive eruptions. Their deposits are found et al. 1982; Garcia et al. 2006). to great depths in all the world’s oceans, but eruptions are typically described by analogy Here we describe the southern cone on the to a subaerial nomenclature that ignores the substantial and inevitable infl uences of hydro- southeast summit plateau of Loihi (18°54′N, static pressure and magma-water interaction at submerged edifi ces. Here we explore mag- 155°15′W), examined in October 2006 with matic volatile exsolution and magma-water interaction for a pyroclastic cone-forming erup- the Hawaiian Undersea Research Laboratory’s tion at ~1 km depth on Loihi Seamount, Hawaii. We examine vesicle textures in lapilli—the Pisces IV submersible. The cone is ~60 m high, physical manifestation of degassing; dissolved volatiles in matrix glasses and olivine-hosted 4 × 106 m3 in volume, with a faintly discernable glass inclusions—the geochemical record of ascent and volatile exsolution; and fi ne ash summit rim we interpret as the edge of a partly morphology—the evidence for if and how external water assisted in fragmentation. This infi lled crater at ~1080 mbsl (Fig. 1C). We col- approach allows a submarine explosive eruption style to be defi ned: the magma achieved lected fi ve large samples of lapilli (2–64 mm)– ~40% vesicularity through almost perfectly closed-system volatile exsolution from ~3 km dominated material from ~1292–1100 mbsl on below the vent, which accelerated and weakened the melt, allowing it to be fragmented by the cone fl anks (Fig. 1C), on which occasional explosive magma-water interaction. We introduce the name “Poseidic” for this end-member subspherical bombs as large as 40 cm were style of submarine basalt explosivity. Poseidic eruptions are identifi able from measurable observed, but not collected intact. Our several- features in pyroclasts, and are possible at all subaqueous basaltic volcanoes. kilogram samples are unique in their volume, grain-size range, and fi eld characterization, INTRODUCTION sive eruption style, here termed “Poseidic,” is and allow us to use measurable features from Explosive submarine eruptions have been the fi rst to be analytically defi ned. a range of the pyroclasts to rigorously interpret described by analogy to subaerial eruption eruption style. styles for which deposit dispersal is a defi n- LOCATION, DEPOSIT, AND METHODS The bulk densities of 225 lapilli were mea- ing characteristic (Walker, 1973), with later Loihi Seamount is the youngest Hawaiian sured following the technique of Houghton and modifi cations to the subaerial classifi cation volcano. Located ~35 km southeast of the island Wilson (1989), who included error estimates scheme incorporating varying mechanisms of Hawaii (Fig. 1A), it rises ~3–5 km from the for such measurements. Vesicle textures were of magma ascent and degassing (Wilson and seafl oor, to an ~12 km2 summit plateau at ~1200 examined on standard polished thin sections. Head, 1981; Vergniolle and Jaupart, 1986; m below sea level (mbsl) (Fig. 1B). The summit Major element glass compositions of lapilli, Parfi tt, 2004; Houghton and Gonnermann, plateau has several conical landforms that reach ash (glass selvages on olivine crystals sieved 2008). Most researchers have interpreted sub- just under 1 km depth, and three pit craters, one directly from the samples), and olivine-hosted marine explosive eruptions as Strombolian of which formed in the most recent eruption glass inclusions, were determined by quantita-

(Davis and Clague, 2006; Clague et al., 2008; tive energy-dispersive electron microprobe. H2O

Sohn et al., 2008), driven by the accumulation and CO2 contents were determined by transmis- and escape of volatiles at unknown rates, from sion Fourier transform infrared spectroscopy 158oW 156oW N S N magma bodies of unconstrained sizes, at inde- A B 0 (FTIR) on lapilli glasses, and by secondary ion Hawaii terminate depths. Such scenarios are diffi cult, 21oN mass spectroscopy (SIMS) on ash and glass Kilauea 4 Loihi if not impossible, to test. Deposits have been Pacific inclusions. FTIR and SIMS have been shown Ocean Mauna Mauna km below sea level Loa characterized from small, almost exclusively o Loihi 8 to yield comparable results for volatile analysis 19 N Loa Pacific ocean crust fi ne-grained samples (Clague et al., 2003, C Summit plateau -155.15.0 (Hauri et al., 2002). Analytical operating con- 2008; Davis and Clague, 2006; Sohn et al., ditions are given in the GSA Data Repository.l 2008), and are often from unidentifi ed vents. -1200 Fine ash particles were examined and character- -1100

These provide little textural, and no dispersal, 18.54.5 ized using scanning electron microscopy. evidence to support analogies with subaerial eruptions of any type. Here we show that at Fault? VESICLE TEXTURES Loihi Seamount, Hawaii, early closed-system Southern cone lapilli have modal bulk vesicu- -1100 degassing into small vesicles that remained cone larity of ~42%, higher than typical Loihi lavas Southern mechanically coupled to their host melt (vol- 100 m atile-coupled degassing) facilitated later interaction of the ascending melt with seawater to Figure 1. A: Hawaiian Islands, with Mauna 1GSA Data Repository item 2010081, supplemen- drive violent eruptions. This submarine explo- Loa, Kilauea, and Loihi volcanoes. B: Loihi tary information and Tables DR1–DR3, is available cross section (Garcia et al., 2006), 2× vertical online at www.geosociety.org/pubs/ft2010.htm, or exaggeration. C: Southern cone bathymetry on request from [email protected] or Docu- (created by J.R. Smith), 20 m contours. Stars ments Secretary, GSA, P.O. Box 9140, Boulder, CO *E-mail: [email protected]. mark sample locations. 80301, USA.

© 2010 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. GEOLOGY,Geology, April April 2010; 2010 v. 38; no. 4; p. 291–294; doi: 10.1130/G30351.1; 4 ﬁ gures; Data Repository item 2010081. 291 Downloaded from geology.gsapubs.org on October 5, 2010

(mostly 0%–20%, reaching ~40% only in rare mon in subaerial Strombolian ejecta (e.g., Lau- (Figs. 3A and 3B). Because the chemistry pillow lavas; Moore et al., 1982; Garcia et al., tze and Houghton, 2005; Polacci et al., 2006). of lapilli and crystals sieved directly from the 2006), and overlapping the range for subaerial deposit does not vary with sample site, the exso- Hawaiian (45%–95%; Cashman and Mangan, VOLATILE EXSOLUTION PROCESSES lution paths captured represent volatile system- 1994) and Strombolian (40%–76%; Lautze and The ~1 km depth of Loihi’s summit is of atics in the southern cone magma as a whole, not Houghton, 2005; Polacci et al., 2006) pyro- interest for volatile exsolution dynamics. At just in an isolated parcel of magma preserved as clasts (Fig. 2A). Quenched sideromelane rims this ~10 MPa pressure, CO2 has limited solu- a single olivine-bearing clast or lava sample. on the Loihi clasts have ~34% subspherical to bility in basalt (Dixon and Stolper, 1995), but A plot of CO2 versus H2O (Fig. 3C) shows polylobate vesicles. Tachylite interiors have H2O solubility is controlled in large part by the that the matrix glasses and glass inclusions ﬁ t a

~45% vesicles, similarly shaped, but with an CO2-H2O-magma system evolution. If exsolved closed-system, volatile-coupled exsolution path additional subpopulation of small subspherical CO2-rich fl uids remain in contact with the melt, from the most CO2-rich inclusion measured vesicles (Fig. 2A). No signifi cant vesicle coales- they thermodynamically stabilize CO2+H2O fl u- (Holloway, 1976; Dixon and Stolper, 1995; cence is apparent. Vesicularity is much less than ids, facilitating subsequent exsolution of water Newman and Lowenstern, 2002); but not an 70%, and vesicles are not signifi cantly intercon- (Holloway, 1976; Dixon and Stolper, 1995; open-system, volatile-decoupled path. The spe- nected, indicating that the magma had limited Dixon et al., 1995; Dixon and Clague, 2001; cial case of volatile decoupling, with CO2-rich permeability (Polacci et al., 2006; Namiki Newman and Lowenstern, 2002), so that vola- magmatic fl uids fl uxing through the magma to and Manga, 2008). Vesicle textures in Loihi tile coupling greatly infl uences eruption style. induce H2O exsolution without the development lapilli resemble those in scoria from the 1959 Lapilli and ash matrix glasses are geochemi- of vesicles (Rust et al., 2004; Spilliaert et al., Kilauea Iki Hawaiian fi re fountains (Fig. 2B). cally identical (Fig. 3A), indicating that the 2006), is unlikely for the southern cone melt Observed episodicity of Kilauea Iki fountains southern cone was derived from a single batch because degassing trends in Figure 3C crosscut implies some degassing variability (Houghton of magma. Southern cone glass inclusions are vapor isopleths, and vesicle textures suggest that and Gonnermann, 2008), but late-stage exsolu- subspherical, hosted in slightly zoned (Fo86–85) the magma permeability required for such fl ux- tion of magmatic H2O in this classic Hawaiian olivine crystals. Matrix glass compositions can ing would have been negligible (Polacci et al., fountaining eruption was melt-coupled (Parfi tt, be obtained by <15% olivine fractionation from 2006; Namiki and Manga, 2008). Our data indi-

2004), producing many small, spherical vesicles glass inclusion compositions, with scatter in cate that: (1) both CO2 and H2O exsolved and in its ejecta. We infer similarly strong volatile the inclusion compositions reﬂ ecting variable contributed to vesiculation; (2) bubbles were coupling for the southern cone’s magma, but post-entrapment crystallization of Cr-spinel insufﬁ ciently interconnected for permeable gas note that Hawaiian eruptions are not unique in having volatiles remain coupled; the same is true SiO (wt%) SiO (wt%) of more-vigorous Plinian eruptions (Houghton 43 45 2 49 51 43 45 2 49 51 and Gonnermann, 2008). Loihi clasts lack the 5 ABD500 large complex-shaped bubbles (Fig. 2C) com- Alk. 80 4 Thol. Cr-Spl Hydromagmatic Ol O (wt%) fragmentation 2 Cr-Spl Ol surface Mg#

O+K 3 Vesicularity (%) 2 60 A Poseidic (Loihi) 75 50 25 0 Na 1500 Hawaiian Loihi lavas Strombolian 750 25 225 clasts 2 C 80 MPa % of clasts % CO 0 Glass inclusions 1.0 2.0 90 Density (g/cm3) Matrix glass (ash) Matrix glass (lapilli) 500 Closed/

coupled Open/ m below sea level (1 wt% initial decoupled ~0.5 cm 1 mm (ppm) 2 vapor)

B Hawaiian C Strombolian CO 3500 250 2 70% CO Volatile-coupled ascent Volatile-coupled

10 MPa

0 8–9 km H O (wt%) 0.20.42 0.81.0

Figure 3. Exsolution. A and B: Total alkali and magnesium number (Mg#) versus SiO2. Alk— ~0.5 cm ~0.5 cm alkalic; Thol—tholeiitic. Fractionation trends for: ~15% olivine (Ol) and ~15% Cr-spinel (Cr-

Spl). C: CO2 versus H2O, with 80 and 10 MPa isobars, 90 and 70 wt% CO2 vapor isopleths, Figure 2. Vesicle textures in scanned thin and closed- and open-system exsolution pathways from the most CO2-rich glass inclusions sections. A: Poseidic Loihi scoria. Density (Newman and Lowenstern, 2002). Closed-system curve requires starting condition of 1 wt% histogram and scanning electron micro- exsolved vapor with >90% CO2 for best ﬁ t. D: Volatile equilibration depths on schematic scope image (phenocryst in gray) inset. southern cone proﬁ le. Analyses and methods are provided in the GSA Repository (see foot- B: Hawaiian scoria. C: Strombolian scoria. note 1).

292 GEOLOGY, April 2010 Downloaded from geology.gsapubs.org on October 5, 2010 loss or fl uxing; and (3) the magma rose quickly level ~200 m below the vent (Fig. 3D), where enough to prevent buoyant bubble decoupling. explosions would have signifi cantly enhanced This scenario is consistent with observed vesi- particle dispersion, driving high eruptive jets cle textures, and inconsistent with Strombolian that explain lapilli equilibration pressures well degassing, which is characterized by open-sys- above the vent (Head and Wilson, 2003). Sub- tem decoupling of gas slugs. It is partly analo- sequent quenching of particles in water then gous to Hawaiian or Plinian degassing, but with prevented welding after deposition. a critical difference imposed by hydrostatic pressure. Magma at Kilauea volcano that loses DISCUSSION the bulk of its CO2 by open-system degassing Subaerial Strombolian bursts require vola- in a summit reservoir may still fountain due to tile decoupling from relatively stagnant melt, volatile-coupled exsolution of magmatic H2O whereas Hawaiian fountains and Plinian jets are alone (Gerlach and Graeber, 1985), whereas driven by volatile-coupled exsolution in more at Loihi, coupled degassing of CO2 and H2O is rapidly ascending melt (Wilson and Head, 1981; required, from at least several kilometers deep Parfi tt, 2004; Houghton and Gonnermann, in the magmatic system. 2008). There is a natural continuum between Equilibrium entrapment pressures of south- these eruption styles: one often passes into the ern cone glass inclusions, calculated from other over time, degassing may be complicated volatile solubility (Dixon and Stolper, 1995; by permeable gas loss and/or post-fragmenta- Newman and Lowenstern, 2002), range from tion vesiculation (Mangan et al., 1993; Cash- 16 to 78 MPa, and indicate progressive entrap- man and Mangan, 1994; Mangan and Cashman, ment by crystals growing in an ascending melt 1996; Polacci et al., 2006; Houghton and Gon- (Fig. 3D). Magma between the 8–9 km shallow Figure 4. Fine ash morphology. A: Poseidic nermann, 2008), and volatile fl uxes measured at Loihi active particles. B: Hawaiian fi ne ash. storage zone at Loihi (Garcia et al., 2006) and subaerial volcanoes commonly confi rm decou- the deepest (~3 km, lithostatic) inclusion ana- pling of a deep-exsolved gas fraction, even in lyzed is not represented in the data. Ash (glass sustained eruptions (Allard et al., 2005). Vesicle selvages) equilibrated to 15–16 MPa, ~200 ± tion resulting from rapid overpressurization of textures and volatile measurements indicate 20 m (lithostatic) below the southern cone vent, bubbles (Spieler et al., 2004) or inertial effects that southern cone Loihi magma ascended and while lapilli equilibrated to pressures of only (Namiki and Manga, 2008) is possible, but degassed in a near-perfect volatile-coupled sys-

5–7 MPa, ~500 ± 100 m (hydrostatic) above the unlikely. The CO2 in vesicles at >7.3 MPa (~730 tem and can be taken, from an in-conduit per- vent. Ash pressures probably represent quench- mbsl) is a relatively dense supercritical fl uid spective, to represent an end-member of volatile ing of effi ciently fragmented melt at the frag- that would accelerate melt less quickly than the behavior in a system more coupled than even mentation surface, with lapilli glasses recording H2O-dominated vapor in vesicles at a subaerial type examples of Hawaiian fi re fountaining. additional volatile exsolution above the vent in vent (Papale and Polacci, 1999); the low viscos- The traditional view that hydrostatic pressure the eruptive jet (Head and Wilson, 2003). ity of basaltic melt allows rapid bubble expan- at submarine vents precludes effi cient volatile sion without fragmentation. Uncoalesced vesi- exsolution and limits explosive potential refl ects FRAGMENTATION cles preserved in lapilli (Fig. 2A) did not burst, the observation that lavas are volumetrically Fragmentation ultimately differentiates an and there are no textural indicators of magmatic dominant on the seafl oor (Moore et al., 1982; effusive from an explosive eruption, and evi- fragmentation in the southern cone ash particles. Garcia et al., 2006). Strong coupling of vola- dence for specifi c mechanisms of fragmenta- Unlike fi ne ash grains, coarse lapilli in Loi- tile-phase bubbles with magma, as captured tion is preserved in the morphology of fi ne ash hi’s southern cone (as in Fig. 2A) are morpho- in this study’s pyroclasts, may be a necessary (Büttner et al., 2002; Zimanowski et al. 2003). logically nondiagnostic of fragmentation mech- precondition to effi cient vesiculation that helps Figure 4 compares blocky Loihi ash particles anism. They have incomplete sideromelane facilitate explosive magma-water fragmentation having scallop-stepped surface fractures, rims and are subequant and fracture bounded, (Trigila et al., 2007) and propel magma to drive adhering microparticles, and surface pitting, suggesting derivation by breakup of decimeter- vigorous, cone-forming submarine explosive with ragged Kilauea Iki ash particles having scale bombs, like those that are visibly intact in eruptions of basalt at depth. Without seawater shapes defi ned by smooth vesicle walls. The the deposits but too brittle to recover. We infer for hydromagmatic fragmentation, the southern Loihi grains are typical hydromagmatic “active that initial in-conduit weakening of the melt cone magma would still likely have erupted in particles” (Büttner et al., 2002), diagnostic of due to magma vesiculation and rapid accelera- vigorous effusion but without pervasive frag- highly energetic thermohydraulic explosions tion was amplifi ed into wholesale fragmenta- mentation (Zimanowski and Büttner, 2003). (Büttner et al., 2002; Zimanowski and Büttner, tion by additional mechanical energy released This eruption style, which we term “Poseidic,” 2003), while Kilauea’s grains are diagnostic of from the localized thermohydraulic explosions accounts for both hydrostatic pressure con- magmatic fragmentation. The active particles that created the active particles. This proceeded straints to degassing, and specifi c effects of are of critical signifi cance. They are formed by a combination of direct shock effects, accel- magma-water interaction. by the most effi cient mechanism of thermal-to- erations away from explosion sites, and hydro- mechanical energy transfer known in volcanic fracturing by steam or water driven through the CONCLUSIONS systems (Zimanowski et al., 2003), and a small magma (Austin-Erickson et al., 2008). Thermo- Submarine pyroclastic deposits must be proportion of them (<10% of ash grains) in the hydraulic explosions did not simply overprint treated as distinct from subaerial ones. Com- deposit represents the dominant fragmentation otherwise fountaining magma; they occurred parisons to subaerial eruptions are useful, but mechanism. below the vent level, initiating and dominat- have limitations: hydrostatic pressure limits At ~42% vesicularity, and with low perme- ing fragmentation. Equilibration pressures of volatile exsolution, interaction of magma with ability, a magmatic contribution to fragmenta- fi ne ash defi ne a hydromagmatic fragmentation seawater is inevitable, and eruption into deep

GEOLOGY, April 2010 293 Downloaded from geology.gsapubs.org on October 5, 2010 water dramatically reduces vent-exit decom- Dixon, J.E., and Clague, D.A., 2001, Volatiles in Newman, S., and Lowenstern, J.B., 2002, Volatile- basaltic glasses from Loihi Seamount, Hawaii: Calc: A silicate melt-H O–CO solution model pression. Together, these produce nonlinear 2 2 differences in degassing history and fragmenta- Evidence for a relatively dry plume compo- written in Visual Basic for Excel: Computers & nent: Journal of Petrology, v. 42, p. 627–654, Geosciences, v. 28, p. 597–604, doi: 10.1016/ tion, dominate plume dynamics, and infl uence doi: 10.1093/petrology/42.3.627. S0098-3004(01)00081-4. particle transport and dispersal in uniquely Dixon, J.E., and Stolper, E.M., 1995, An experimen- Papale, P., and Polacci, M., 1999, Role of carbon diox- submarine ways. The Poseidic eruption style is tal study of water and carbon dioxide solubili- ide in the dynamics of magma ascent in explo- the fi rst to be characterized fully on the basis ties in mid-ocean ridge basaltic liquids, Part II: sive eruptions: Bulletin of Volcanology, v. 60, Applications to degassing: Journal of Petrol- p. 583–594, doi: 10.1007/s004450050253. of measurable deposit and pyroclast features. ogy, v. 36, p. 1633–1646. Parfi tt, E.A., 2004, A discussion of the mechanisms More work, especially on pyroclastic deposits Dixon, J.E., Stolper, E.M., and Holloway, J.R., 1995, of explosive basaltic eruptions: Journal of Volca- below ~1 km, is required to assess the role of An experimental study of water and carbon di- nology and Geothermal Research, v. 134, p. 77– Poseidic eruptions in the deep ocean, and the oxide solubilities in mid-ocean ridge basaltic 107, doi: 10.1016/j.jvolgeores.2004.01.002. extent to which other submarine pyroclastic liquids, Part I: Calibration and solubility mod- Polacci, M., Corsaro, R.A., and Andronico, D., 2006, els: Journal of Petrology, v. 36, p. 1607–1631. Coupled textural and compositional charac- eruptions are driven by decoupled rather than Garcia, M.O., Caplan-Auerbach, J., De Carlo, E.H., terization of basaltic scoria; Insights into the coupled volatile volumes. Poseidic eruptions Kurz, M.D., and Becker, N., 2006, Geology, transition from Strombolian to fi re fountain are possible at any subaqueous basaltic vol- geochemistry and earthquake history of Loihi activity at Mount Etna, Italy: Geology, v. 34, cano. Because these are the most abundant Seamount, Hawaii’s youngest volcano: Chemie p. 201–204, doi: 10.1130/G22318.1. volcanoes on Earth, we expect that Poseidic der Erde Geochemistry, v. 66, p. 81–108, doi: Rust, A.C., Cashman, K.V., and Wallace, P., 2004, 10.1016/j.chemer.2005.09.002. Magma degassing buffered by vapor fl ow eruption products will be recognized in many Gerlach, T.M., and Graeber, E.J., 1985, Volatile through brecciated conduit margins: Geology, seafl oor sequences worldwide. budget of Kilauea Volcano: Nature, v. 313, v. 32, p. 349–352, doi: 10.1130/G20388.2. p. 273–277. Sohn, R.A., et al., 2008, Explosive volcanism on ACKNOWLEDGMENTS Hauri, E., Wang, J., Dixon, J.E., King, P.L., Mande- the ultraslow-spreading Gakkel ridge, Arc- T. Kirby, J.R. Smith, and the Hawaii Undersea Re- ville, C., and Newman, S., 2002, SIMS analy- tic Ocean: Nature, v. 453, p. 1236–1238, doi: search Laboratory ensured successful dives; H. Gon- sis of volatiles in silicate glasses 1: Calibration, 10.1038/nature07075. nermann, G. Valentine, and H. Tuffen reviewed earlier matrix effects and comparison with FTIR: Spieler, O., Kennedy, B., Kueppers, U., Dingwell, manuscripts; W.K. Stovall provided Kilauea Iki sam- Chemical Geology, v. 183, p. 99–114. D.B., Scheu, B., and Taddeucci, J., 2004, The ples. Dives were funded by the National Oceanic and Head, J.W.I., and Wilson, L., 2003, Deep submarine fragmentation threshold of pyroclastic rocks: Atmospheric Administration; post-dive support was pyroclastic eruptions: theory and predicted Earth and Planetary Science Letters, v. 226, from the University of Otago and the Foundation for landforms and deposits: Journal of Volcanol- p. 139–148, doi: 10.1016/j.epsl.2004.07.016. Research and Technology, New Zealand. ogy and Geothermal Research, v. 121, p. 155– Spilliaert, N., Allard, P., Metrich, N., and Sobolev, 193, doi: 10.1016/S0377-0273(02)00425-0. A.V., 2006, Melt inclusion record of the con- REFERENCES CITED Holloway, J.R., 1976, Fluids in the evolution of ditions of ascent, degassing and extrusion of Allard, P., Burton, M., and Mure, F., 2005, Spectro- granitic magmas: consequences of fi nite CO2 volatile-rich alkali basalt during the powerful scopic evidence for a lava fountain driven by solubility: Geological Society of America 2002 fl ank eruption of Mount Etna (Italy): previously accumulated magmatic gas: Nature, Bulletin, v. 87, p. 1513–1518, doi: 10.1130/ Journal of Geophysical Research, v. 111, v. 433, p. 407–410, doi: 10.1038/nature03246. 0016-7606(1976)87<1513:FITEOG>2.0.CO;2. B04203, doi: 10.1029/2005JB003934. Austin-Erickson, A., Büttner, R., Dellino, P., Ort, Houghton, B.F., and Gonnermann, H.M., 2008, Ba- Trigila, R., Battaglia, M., and Manga, M., 2007, An M.H., and Zimanowski, B., 2008, Phre- saltic explosive volcanism: Constraints from experimental facility for investigating hydro- atomagmatic explosions of rhyolitic magma: deposits and models: Chemie der Erde Geo- magmatic eruptions at high-pressure and high- Experimental and fi eld evidence: Journal of chemistry, v. 68, p. 117–140, doi: 10.1016/j. temperature with application to the importance Geophysical Research, v. 113, B11201, doi: chemer.2008.04.002. of magma porosity for magma-water interac- 10.1029/2008JB005731. Houghton, B.F., and Wilson, C.J.N., 1989, A vesicu- tion: Bulletin of Volcanology, v. 69, p. 365– Büttner, R., Dellino, P., La Volpe, L., Lorenz, V., and larity index for pyroclastic deposits: Bulletin of 372, doi: 10.1007/s00445-006-0081-6. Zimanowski, B., 2002, Thermohydraulic explo- Volcanology, v. 51, p. 451–462, doi: 10.1007/ Vergniolle, S., and Jaupart, C., 1986, Separated two- sions in phreatomagmatic eruptions as evidenced BF01078811. phase fl ow and basaltic eruptions: Journal of by the comparison between pyroclasts and prod- Lautze, N.C., and Houghton, B.F., 2005, Physical Geophysical Research, v. 91, p. 12842–12860, ucts from Molten Fuel Coolant Interaction exper- mingling of magma and complex eruption doi: 10.1029/JB091iB12p12842. iments: Journal of Geophysical Research, v.107, dynamics in the shallow conduit at Stromboli Walker, G.P.L., 1973, Explosive volcanic eruptions: n. B11, 2277, doi: 10.1029/2001JB000511. volcano, Italy: Geology, v. 33, p. 425–428, doi: A new classifi cation scheme: Geologische Cashman, K.V., and Mangan, M.T., 1994, Physical 10.1130/G21325.1. Rundschau, v. 62, p. 431–446, doi: 10.1007/ aspects of magmatic degassing; II, Constraints Mangan, M.T., and Cashman, K.V., 1996, The struc- BF01840108. on vesiculation processes from textural studies ture of basaltic scoria and reticulite and infer- Wilson, L., and Head, J.W., 1981, Ascent and em- of eruptive products: Reviews in Mineralogy ences for vesiculation, foam formation, and placement of basaltic magma on the Earth and and Geochemistry, v. 30, p. 447–478. fragmentation in lava fountains: Journal of Moon: Journal of Geophysical Research, v. 86, Clague, D.A., Batiza, R., Head, J.W.I., and Davis, Volcanology and Geothermal Research, v. 73, p. 2971–3001, doi: 10.1029/JB086iB04p02971. A.S., 2003, Pyroclastic and hydroclastic de- p. 1–18, doi: 10.1016/0377-0273(96)00018-2. Zimanowski, B., and Büttner, R., 2003, Phreatomag- posits on Loihi Seamount, Hawaii, in White, Mangan, M.T., Cashman, K.V., and Newman, S., 1993, matic explosions in subaqueous volcanism, in J.D.L., et al., eds., Explosive subaqueous vol- Vesiculation of basaltic magma during erup- White, J.D.L., et al., eds., Explosive Subaque- canism: American Geophysical Union Mono- tion: Geology, v. 21, p. 157–160, doi: 10.1130/ ous Volcanism: American Geophysical Union graph 140, p. 73–95. 0091-7613(1993)021<0157:VOBMDE>2.3 Monograph 140, p. 51–60. Clague, D.A., Paduan, J.B., and Davis, A.S., 2008, .CO;2. Zimanowski, B., Wohletz, K., Dellino, P., and Büttner, Widespread strombolian eruptions of mid- Moore, J.G., Clague, D.A., and Normark, W.R., 1982, R., 2003, The volcanic ash problem: Journal of ocean ridge basalt: Journal of Volcanology and Diverse basalt types from Loihi Seamount, Ha- Volcanology and Geothermal Research, v. 122, Geothermal Research, v. 180, p. 171–188, doi: waii: Geology, v. 10, p. 88–92, doi: 10.1130/ p. 1–5, doi: 10.1016/S0377-0273(02)00471-7. 10.1016/j.jvolgeores.2008.08.007. 0091-7613(1982)10<88:DBTFLS>2.0.CO;2. Davis, A.S., and Clague, D.A., 2006, Volcaniclastic Namiki, A., and Manga, M., 2008, Transition be- Manuscript received 21 April 2009 deposits from the North Arch volcanic fi eld, Ha- tween fragmentation and permeable outgassing Revised manuscript received 30 September 2009 waii: Explosive fragmentation of alkalic lava at of low viscosity magmas: Journal of Volcanol- Manuscript accepted 14 October 2009 abyssal depths: Bulletin of Volcanology, v. 68, ogy and Geothermal Research, v. 169, p. 48– p. 294–307, doi: 10.1007/s00445-005-0008-7. 60, doi: 10.1016/j.jvolgeores.2007.07.020. Printed in USA

294 GEOLOGY, April 2010