Beruflich Dokumente
Kultur Dokumente
Sarah A. Fagents
University of Hawaii
Tracy K. P. Gregg
University at Buffalo
Rosaly M. C. Lopes
Jet Propulsion Laboratory
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521895439
A catalog record for this publication is available from the British Library
Just as we were putting the finishing touches on the field with him, from scrambling up impos-
this book, we learnt of the unexpected passing sibly loose cinder cones in the high desert of
of Professor Ronald Greeley, Regents’ Professor Arizona, to some eventful wilderness camping
in the School of Earth and Space Exploration at in the Pinacate Volcanic Field in Sonora, Mexico.
Arizona State University. Ron was a friend and Here his playful side would come to the fore,
mentor to all three of us editors and a pioneer and we would all learn as much as we had fun.
in the study of planetary geology and volcan- Tracy met Ron at the insistence of Dr.
ology. He was involved in lunar and planetary Peter Schultz, who was Tracy’s undergradu-
exploration since 1967 and contributed signifi- ate research advisor at Brown University. Pete
cantly to our understanding of the geology of asserted that Ron’s experience and knowledge
the Moon, Mars, Europa, and several other Solar in planetary volcanism was a perfect match for
System bodies. He participated in nearly every Tracy’s graduate school interests. Ron welcomed
major space mission flown in the Solar System Tracy into his research group at Arizona State
since the Apollo Moon landing. Above all, Ron University when she arrived to begin her gradu-
was a superb teacher and mentor who touched ate program in 1990. Although Tracy ultimately
many lives and enabled many careers in planet- received her PhD under the advisement of Dr.
ary science. His reach was extensive and it is not Jon Fink, Ron coached her through the writ-
surprising that all three of us learnt much from ing and publication of her first peer-reviewed
him, although we came from different scientific journal article, and she still uses the lessons he
backgrounds and even different countries. taught her about scientific writing today: Ron’s
Ron hired Sarah as a post-doc in 1996, shortly words and phrases literally flow through her fin-
after she had moved to the United States from gers onto the pages of this very book. After Tracy
the United Kingdom. He provided unparalleled became a professor at the University at Buffalo,
mentorship as she bridged the chasm between she found herself turning to Ron frequently for
wide-eyed graduate student and fully fledged professional and diplomatic advice in navigat-
independent research scientist. Ron challenged ing the waters of proposals, papers, and grants.
all his students and post-docs to really stretch His experience was invaluable and will be sorely
themselves. His in-at-the-deep end approach missed.
introduced Sarah to whole new research areas, Rosaly first met Ron while she was start-
from lava–ice interactions in Iceland and on ing her graduate studies at University College
Mars, to the icy satellites of Jupiter. His gener- London. Her advisor, John Guest, was a close
osity with research, fieldwork, and travel oppor- friend of Ron’s and the two of them did field
tunities allowed her to engage with research work together on several volcanoes around the
scientists around the globe, and to gain those world. Rosaly joined Ron and John in Hawaii in
vital first footholds en route to becoming estab- 1981 to help them with field work on Kilauea.
lished in the scientific community. She attributes Coming from a physics and astronomy back-
her subsequent research successes to Ron’s ground, she had much to learn about geology
exquisite, though sometimes painful, tutelage field work and benefited not only from Ron’s
in the fine art of proposal-writing. Perhaps the expertise but also his patience and true gift for
fondest memories she has of Ron are of being in teaching. She had been warned about Ron and
Index 414
Color Plates to be found between pages x and y.
Costanza Bonadonna, Section des Sciences University of Hawaii, 73–4460 Queen KaahuÂ�
de la Terre et de l’Environnement, Université manu Hwy., #119, Kailua-Kona, HI 96740; (808)
de Genève, 13 rue des Maraîchers, CH-1205 327 6206, fax: (808) 327 6207.
Genève, Switzerland; +41 22 379 3055; fax +41
22 379 3210. Helge M. Gonnermann, Department of Earth
Science, Rice University, MS 126, 6100 Main
Ralf Büttner, Physikalisch Vulkanologisches Street, Houston, Texas, 77005; (713) 348 6263,
Labor, Universität Würzburg, Pleicherwall 1, fax (713) 348 5214.
Würzburg, D-97070, Germany; +49 (0)931 318
2387. Tracy K. P. Gregg, Department of Geology, 411
Cooke Hall, University at Buffalo, Buffalo, NY
Bernard Chouet, United States Geological 14260; (716) 645 4328, fax (716) 645 3999.
Survey, Volcano Science Center, 345 Middlefield
Road, MS 910, Menlo Park, CA 94025; (650) 329 Andrew J. L. Harris, Laboratoire Magmas et
4796, fax (650) 329 5203. Volcans, Université Blaise Pascal, 5 rue Kessler,
63038 Clermont-Ferrand, France; +33 (0)4 73 34
Antonio Costa, Osservatorio Vesuviano, INGV, 67 37, fax +33 (0)4 73 34 67 44.
Napoli, Italy, and Environmental Systems
Science Centre, University of Reading, Reading James W. Head, Department of Geological
RG6 6AL, UK. +44 (0)118 378 8741, fax +44 (0)118 Sciences, Brown University, 324 Brook St.,
378 6413. Providence, RI 02912; (401) 863 2526, fax (401)
863–3978.
Amanda B. Clarke, School of Earth and Space
Exploration, Arizona State University, Box Bruce F. Houghton, Department of Geology
871404, Tempe, AZ 85287; (480) 965–6590, fax and Geophysics, University of Hawaii at Manoa,
(480) 965–8102. 1680 East-West Road, Honolulu, HI 96822; (808)
956 2561, fax (808) 956 5512.
Josef Dufek, School of Earth and Atmospheric
Science, Georgia Institute of Technology, 311 Christian Huber, School of Earth and
Ferst Drive, Atlanta, GA 30332. Atmospheric Science, Georgia Institute of
Technology, 311 Ferst Drive, Atlanta, GA 30332;
Sarah A. Fagents, Hawaii Institute of Geo� (404) 385 1509, fax (404) 894 5638.
physics and Planetology, University of Hawaii
at Manoa, 1680 East-West Road, Honolulu, HI Mike James, Lancaster Environment Center,
96822; (808) 956 3163, fax (808) 956 6322. Lancaster University, Lancaster LA1 4YQ, UK;
+44 (0)1524 593571, fax +44 (0)1524 593985.
David Fee, Wilson Infrasound Observatories,
Geophysical Institute, University of Alaska Leif Karlstrom, Department of Earth and
Fairbanks, 903 Koyukuk Drive, Fairbanks, Planetary Science, 307 McCone Hall, University
AK 99775–7320; (907) 474 7564, fax (907) 474 of California, Berkeley, Berkeley, CA 94720; (510)
7290. 642–3993, fax (510) 643–9980.
5 rue Kessler, 63038 Clermont-Ferrand, France; Olivier Roche, Laboratoire Magmas et Volcans,
+33 (0)4 73 34 67 41, fax +33 (0)4 73 34 67 44. Université Blaise Pascal, CNRS, IRD, 5 rue Kessler,
63038 Clermont-Ferrand, France; +33 (0)4 73 34
Steve J. Lane, Lancaster Environment Center, 67 68, fax +33 (0)4 73 34 67 44.
Lancaster University, Lancaster LA1 4YQ, UK;
+44 (0)1524 593437, fax +44 (0)1524 593985. John L. Smellie *, British Antarctic Survey, High
Cross, Madingley Road, Cambridge CB3 0ET, UK.
Rosaly M. C. Lopes, Earth and Space Sciences *Now at: Department of Geology, University of
Division, Mail Stop 183–601, Jet Propulsion Leicester, University Road, Leicester LE1 7RH,
Laboratory, Pasadena, CA 91109; (818) 393 4584, UK; +44 (0)116 252 3933, fax +44 (0)116 252
fax (818) 393 3218. 3918.
Michael Manga, Department of Earth and Benoit Taisne, Institut de Physique du Globe
Planetary Science, UC Berkeley, 307 McCone de Paris, 1 rue Jussieu, 75238 Paris Cedex 05,
Hall, Berkeley, CA 94720; (510) 643–8532, fax France; +33 (0)1 83 95 75 79, fax +33 (0)1 83 95
(510) 643–9980. 77 02.
Jon J. Major, US Geological Survey, 1300 SE Steve Tait, Institute de Physique du Globe
Cardinal Court, Bldg 10, Suite 100, Vancouver, de Paris, 1 rue Jussieu, 75238 Paris Cedex 05,
WA 98683; (360) 993 8927 fax: (360) 993 8980. France; +33 (0)1 83 95 74 25; fax +33 (0)1 83 95
77 02.
Vernon Manville, School of Earth and
Environment, University of Leeds, Leeds LS2 Lionel Wilson, Lancaster Environment Center,
9JT, UK; +44 (0)113 3431307, fax +44 (0)113 343 Lancaster University, Lancaster LA1 4YQ, UK;
5259. +44 1524 593889, fax +44 1524 510217.
Robin Matoza, Institute of Geophysics and Ken Wohletz, Geophysics Group, EES-17,
Planetary Physics, Scripps Institution of MS F665, Los Alamos National Laboratory, Los
Oceanography, La Jolla, California, USA. Alamos, NM 87545; (505) 667–9202, fax (505)
665–3687.
Karl L. Mitchell, Jet Propulsion Laboratory,
California Institute of Technology, Mail Stop Andrew W. Woods, BP Institute, University
183–601, 4800 Oak Grove Dr., Pasadena, CA of Cambridge, Bullard Laboratories, Madingley
91109–8099, USA; (818) 393–5519. Rise, Madingley Road, Cambridge CB3 0EZ.
Introduction
Sarah A. Fagents, Tracy K. P. Gregg and Rosaly M. C. Lopes
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
research or professional volcanologist could eruption, and deposition at the surface, fol-
draw on in seeking up-to-date information on lowed by post-deposition modification. We thus
modeling approaches to volcanic phenomena or cover the thermal and dynamic processes tak-
a point of access to a new volcanologic subdis- ing place in accumulations of magma at depth
cipline. While the book therefore has elements in the crust (Chapter 2), before considering
aimed at students, it does not have the trad- the ways in which magma ascends to the sur-
itional textbook style and is rich in references, face (Chapters 3 and 4). Processes taking place
both historic and contemporary. within the conduit, such as the evolution of
The increasing populations living near active magma rheology and the behavior of magmatic
volcanoes imply rising volcanic risk worldwide. (or external) volatiles, determine the subaerial
Therefore, the role of the well-prepared vol- manifestation of the eruption; whether as an
canologist, and the early training thereof, is effusion of lava (Chapter 5), or explosive mani-
becoming increasingly important. Thus this festation such as the unsteady, transient proc-
book is conceived to bridge the gap in the litera- esses of strombolian (Chapter 6) and vulcanian
ture by offering quantitative treatments of vol- explosions (Chapter 7), or the sustained explo-
canic processes, from the subsurface through sive columns of sub-plinian or plinian erup-
deposition and post-eruptive modification, at tions, and lava fountains of hawaiian activity
a level designed for students, their professors, (Chapter 8). Chapter 9 considers the fallout of
and professional volcanologists alike to come tephra ejected in sustained explosive eruptions,
efficiently up to speed on the state of the art in while Chapter 10 considers the emplacement
understanding the physics and mathematics of of material as pyroclastic density currents. The
volcanism. interaction of magma with its environment can
produce a diverse range of eruption styles. For
example surface or near-surface water interact-
Content ing with erupting magma has the potential to
produce violently explosive activity with dis-
This book treats a wide variety of volcanic proc- tinctive eruptive products (Chapter 11), while
esses, from the dynamics of magma accumula- deep-sea eruptions produce a variety of effusive
tion deep beneath the Earth’s surface through and weakly explosive activity that dominates
ascent, eruption, and deposition of volcanic Earth’s volcanic record on a volumetric basis
products. In compiling this volume, we sought (Chapter 12). The interaction of magma with ice
out experts in each subdiscipline of quantita- is considered in Chapter 13, while Chapter 14
tive physical volcanology in order to ensure addresses the syn- or post-eruptive phenomena
that each subject area was treated comprehen- of lahars. Volcano monitoring techniques (i.e.,
sively. Thus, each of the chapters that follow seismology and infrasound) exploit the physics
is devoted to a specific volcanic process and of volcanic processes to elucidate a broad spec-
authored by specialists in that subject matter. trum of volcanic behavior (Chapters 15 and 16).
Each chapter is designed to take the reader from Looking outward from planet Earth, we also
the basic physics through to the state of the art consider how models of terrestrial volcanism
in understanding the volcanic process through can be adapted to help us understand the vol-
modeling, illustrated with field examples. Most canism of other planets (Chapter 17).
chapters include a set of exercises at the end, While we followed this logical order, there
designed to consolidate the reader’s under- are many commonalities among chapters.
standing of the subject matter. For example, the intricacies of and feedbacks
The order of the book chapters logically among volatile exsolution, magma rheologic
follows the sequence of volcanic processes behavior, crystal growth, and multiphase flow
observed in nature: we begin with what hap- are critical to eruptive behavior, and so these
pens at great depths beneath the Earth’s sur- themes are addressed in several chapters. In
face and track the magma through ascent, addition, the dynamics of granular flow are
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
consider the conditions for the formation and in the eruption and emplacement process. The
evolution of magma chambers from thermal, very fact that the magmas are eruptible, at rela-
solid stress, and fluid mechanics perspectives. tively low crystallinity, may suggest that this
Magma chambers are commonly envis- state is not necessarily representative of a large
aged as high melt fraction bodies (> 40% melt). portion of the magmatic history (Marsh, 1981).
However, thermal and dynamic arguments typ- While many difficulties exist in interpret-
ically favor these intrusive bodies existing at ing both the plutonic and volcanic record to
lower melt fractions for considerable propor- infer dynamics, geochemical tracers continue
tions of their histories (Marsh, 1981). Large, to be among the best sources of information
high melt fraction chambers have existed at about the dynamics of these systems (Wyllie,
least transiently in the crust, as evidenced by 1977; Marsh, 1989; Bachmann et al., 2007a).
voluminous ignimbrite sheets and correspond- Importantly, diffusion-based and radioactive
ing caldera collapse structures (Spera and Crisp, decay-based chronometers indicate that phen-
1981; Druitt and Francavigilia, 1992). However, ocrysts in magmatic systems can grow over a
buffering by latent heat release and diminishing range of timescales and can exhibit prolonged
thermal gradients during cooling of a magma histories at suprasolidus temperatures in the
intrusion implies longer periods of time at low crust, sometimes in excess of hundreds of ka
melt fraction. This chapter adopts the definition (Costa et al., 2003; Vazquez, 2004; Simon and
of a magma chamber as a spatially connected Reid, 2005; Reid, 2008). The decay of uranium
body of magma in a suprasolidus state, and series nuclides has also shown that different
uses this term generically regardless of whether crystal populations can have different residence
the chamber is composed mostly of melt or a times in magma chambers, can indicate mix-
crystal-rich mush. ing of complex populations from in-situ crystal-
Speculation on chamber processes dates lization and assimilation from older wall rock
to the recognition of some intrusive suites as (Cooper and Reid, 2003), and can indicate much
being of fluidal, magmatic origin (Hutton, 1788). shorter bursts of volatile exsolution and gas
Consideration of the dynamics of convection transport (Berlo et al., 2006).
in magma chambers dates back to at least the However, these measurements of magmatic
late nineteenth century (Grout, 1918). Modern, residence times do not give precise information
rock-record-based investigations of chamber about the spatial and temporal variability in
dynamics are largely based on two parallel melt fraction or the dynamics of these systems.
sources of data: the examination of exhumed Real-time geophysical observations, though
plutonic rocks and the examination of eruptive still sparse, are beginning to provide import-
products (Bachmann et al., 2007a). Both ana- ant snapshots of the crust and short-duration
lyses have been useful in delineating aspects magmatic system evolution. These investiga-
of chamber processes, although they provide tions are primarily conducted by observing how
very different types of data. The plutonic realm seismic waves interact with the crust and by
provides a time-integrated view of magmatic observing the deformation of the ground due
processes, but reflects different parts of the to the changes in magmatic systems (geodesy).
temporal record with unequal fidelity. Eruptive Gravity and changes in the conductivity have
products tend to provide closer to an instant- also proven to be useful tools to describe mag-
aneous snapshot of the compositional field of a matic systems in certain localities (Ajakaiye,
chamber at the time of the eruption. However, 1970; Manzella et al., 2000; Gudmundsson and
some time-integrative information still exists Hognadottir, 2007).
in the form of phenocrysts that crystallized Seismic tomographic inversion has identi-
and were advected with the magma (Wallace fied low-velocity regions beneath volcanic edi-
and Bergantz, 2005). Interpretation of the com- fices that may correspond to a combination of
positional field from eruptive products can be regions of partial melt and elevated tempera-
somewhat confounded by complex dynamics tures. Nevertheless, the elastic properties of
rock, as shown by laboratory measurements, Modeling also helps us to understand these bod-
depend on several other variables such as tem- ies in the broader context of crustal evolution
perature, pressure, rock type, and the presence and volcano–pluton connections. Section 2.2
of hydrothermal fluids. The non-unique inter- discusses the thermal evolution of the crust in
pretations for spatial variation of seismic vel- magmatically active regions and its relationship
ocities makes it generally difficult to attribute to the evolution of chambers and geochem-
those effects solely to the presence of melt, and ical observations of chamber residence times.
requires good a priori knowledge of the geol- Section 2.3 examines the relationship between
ogy of the region to make accurate estimates magma chambers and crustal stress fields,
of melt fraction (Lees, 2007). Additionally, the and how this relates to geodetic observations.
problem of resolution due to both ray coverage Section 2.4 describes magma motion inside
and the large-scale filtering of fine structures by magma chambers, and the multiphase interac-
seismic wavelengths generally on the order of tions between melt and discrete phases (such as
a few hundred meters to kilometers introduces crystals and bubbles). Section 2.5 is devoted to
blurring of the geometry and size of the magma a discussion of future directions in the physical
body. examination of magma chambers.
The measurement of ground deformation in
volcanic areas provides another observational
indication of the presence of magma cham-
bers. Geodetic studies fit time-series of ground 2.2╇ ╇Heat transfer and magmatic
deformation to models, which are then used intrusions
to back out processes, stresses, and volumes of
intruded magma. Data are collected primarily The magmas that accumulate in crustal magma
by GPS units deployed on the edifice (Segall and chambers are either generated by mantle melt-
Davis, 1997), or by Interferometric Synthetic ing with subsequent crustal transport, or are a
Aperture Radar (InSAR) (Poland et al., 2006). result of crustal-level melting due to anomalous
Although sensitive to rheological variations, thermal conditions. While many crustal melts
including those induced by different melt frac- can be generated in proximity to intrusions that
tions, ground deformation is primarily used to have heated surrounding rocks (Bergantz, 1995),
indicate changes in volume of magma chambers or result from over-thickening of the crust in
due to either inflation or deflation of a cham- orogenic zones (Patiño Douce and Harris, 1998),
ber as a result of magmatic flux. Despite diffi- the advection of mass and heat from the mantle
culties in gathering and interpreting these data, ultimately drives crustal-level magmatism.
a growing database of geodetic observations and After accumulation, the long-term viability
increasing sophistication of inversion meth- of magma chambers in the crust is critically
ods provide unique insight into active magma dependent on the surrounding thermal envir-
chamber dynamics. onment, and the thermodynamic as well as
Diverse observations, from geochemically rheological properties of the melt and country
inferred residence times to remote imaging rock (Patiño Douce and Harris, 1998; Newman
methods, show that subsurface magmatic sys- et al., 2006a; Karlstrom et al., 2009). Relatively
tems are active on a variety of time and length cold country rock will drive large thermal gra-
scales. However, our lack of direct observation dients between the magma body and its mar-
of these systems has limited our ability to under- gins, limit the time a body of magma remains
stand how they evolve dynamically in time. melt-dominated, and also inhibit the partial
Therefore, mathematical, experimental, and melting and assimilation of wall rock material.
computational models are widely used to inte- Likewise, hydrothermal circulation transports
grate and explain some of the disparate obser- thermal energy away from the immediate vicin-
vations and provide a framework that can help ity of an intrusion, leading to enhanced solidi-
guide future observations of magma chambers. fication in the shallow crust. In contrast, all
condition for the heat budget, and in this sense The generalized equation for enthalpy evolu-
is a fully coupled problem. In contrast, a speci- tion takes the form
fied, constant-temperature boundary condition
∂H T ∂ ∂ ∂T
at the margin of a magma chamber would gen- + (u i H T ) = k , (2.4)
erate unphysical amounts of crustal melt. ∂t ∂x i ∂x i ∂x i
The thermal interface between the magma
and wall rock can be described as (Carrigan, where the enthalpy (HT) is composed of a sens-
1988): ible (heat that results in a temperature change)
and latent heat component:
∂T
k = h (Tb − TR ) . (2.1)
∂z S
ensible Heat
Latent Heat
b
T (2.5)
H Τ = ρ∫ c ρ dT + ρ fL .
Tref
All notation is summarized in Section 2.7. Here
Tb is the temperature at the boundary, TR is the Here ρ is the magma density (kg m–3), cp is the
temperature of the thermally unperturbed wall magma specific heat capacity (J kg–1 K–1), L is the
rock some distance from the chamber, k is the latent heat (J kg–1), f is the melt fraction (volu-
thermal conductivity of the magma (W K–1 m–1), metric), and ui is the velocity vector. For the
and h a heat transfer coefficient (W K–1). This case of constant heat capacity and conductivity,
boundary condition relation can be made dimen- normalizing density by a reference density (ρ0),
sionless by introducing z* = z/δ, where δ is the characteristic length scale (δ), velocity (u0), and
thickness of the boundary layer inside the cham- temperature (T0), the heat conservation can be
ber where convective motion ceases, and by set- recast in dimensionless form to give:
ting T* = T/TR. This then results in the equation, ∂T *
∂T * 1 ∂ 2T * R*
ρ* + u *i = − . (2.6)
*
∂t * ∂x i Pe ∂x i*
2
∂T * Ste
= Bi (T * −1). (2.2)
∂z * b
where R* is the dimensionless rate of production
The dimensionless term, the Biot number (Bi), of melt, given by:
gives the ratio of thermal resistance of a magma
∂f
chamber to the thermal resistance of the R * = ρ* . (2.7)
Â�surrounding rock: ∂t *
hδ
Bi = . (2.3) Two dimensionless parameters are introduced
k in this formulation, the Stefan number and the
Peclet number. The Stefan number (Ste) is a ratio
A summary of dimensionless numbers used of the sensible heat to the latent heat contribu-
commonly in magma dynamics is given in tions in the flow, given as
Online Supplement 2A (see end of chapter).
c pT0
Most estimates of magma chamber conditions Ste = . (2.8)
indicate Bi ≪ 1 (Carrigan, 1988). Under these L
conditions heat transfer is limited by the ther-
mal resistance of the wall rock, even in the pres- The Peclet number (Pe) is defined as the ratio of
ence of hydrothermal circulation. Described as advective to diffusive heat transport:
the “thermos-bottle effect”, the amount of heat
c p ρ0 δ
lost into the wall rock is limited by the more Pe = u0 . (2.9)
insulating wall rock. Accordingly, we focus first k
on conduction of heat into the wall rock and the
limitations this places on the thermal evolution In the case of heat transport in the solid crust,
of the magma chamber. the Peclet number is small and we can neglect
Study Model type1 Intr. style2 Total intr.3 (km) Tinit (°C) Rock type4 Tmelt–Tsolid (°C) E (%)
9780521895439c02_p5-31.indd 10
Younker and 1D, cond., no single intrusion 2.0 500 basalt, L:1200, S:1100 32
Vogel, 1976 bottom heat loss biotite-granite L:1100, S:800
Wells, 1980 1D, cond., multiple 40.0 200 tonalit e L:1050, S:800 8
over-accretion intrusion
Huppert and 1D, param. single intrusion 0.5 500 basalt, L:1200, S:1091 44
Sparks, 1988 convection, no grano-diorite L:1000, S:850
bottom heat loss
Bergantz, 1989 1D, cond., no single intrusion 16.6 700 basalt, pelite L:1250, S: 980 38
bottom heat loss L:1200, S: 725
Bittner and 2D, convection single intrusion 5.0 756 basalt, granite L:1100, S: 950 NA
Schmeling, 1995 L:1050, S:760
Barboza and 2D, convection fixed T bottom NA 600 pelite L:1200, S:750 NA
Bergantz, 1996 boundary
Raia and Spera, 2D, convection fixed T bottom NA 1195 (CaAl2Si2O8 - L:1547, S:1277 NA
1997 boundary CaMgSi2O6)
Pedersen et al., 1D, cond., multiple 10.0 650 basalt, L:1250, S:1100 5
1988 over-accretion intrusion grano-diorite L:1000, S:710
Petford and 1D, cond., multiple 1.0 650 basalt. L:1250, S:1050 4
Gallagher, 2001 over-accretion intrusion amphibolite L:1075, S:1010
Annen and 1D, cond, multiple 8.0 variable, based basalt, L:1300, S:620 8
Sparks, 2002 over-accretion intrusion on depth (600) amphibolite L:1075, S:1010
Dufek and 2D, cond. and multiple variable (5.0) variable, based basalt, pressure dependent 0.5–10.4 (7.1)
Bergantz, convection, intrusion on depth (640) amphibolite L:1240, S:640
2005a stochastic L: 1100, S:850
1
╇ Dimension, heat conduction or convection.
2
╇ Intruding magma, physical configuration of intrusion (sill, etc.) or specified temperature boundary condition.
3
╇ Integrated melt volume/basal area, or for 1D models integrated melt height.
4
╇ Intruded magma listed first, then country-rock.
5
╇ Parentheses indicate specific example.
8/2/2012 7:51:51 AM
M AG M A C H A M B E R DY N A M I C S A N D T H E R M O DY N A M I C S 11
Figure 2.3╇╇Three scales of melt–crust interaction. (a) Dike network of 1.43 Ga Vernal Mesa pegmatite cross cutting 1.73 Ga
gneisses on the 670 m high Painted Wall, Black Canyon of the Gunnison, Colorado. Dikes were emplaced at about 10 km depth in
tensional opening during dextral/oblique shear on the Black Canyon shear zone (Jessup et al., 2006). (b) Orthogonal dike network
of ~1.69 Ga granite and pegmatite cross cutting 1.73 Ga granodiorite within Tuna Canyon, tributary to Upper Granite Gorge of
Grand Canyon. Dike networks in Grand Canyon reflect late-syntectonic granite dike swarms emplaced at about 20 km depth
(Dumond et al., 2007) during late stages of accretionary assembly of southwestern North American lithosphere (note geologist
for scale) (Ilg et al., 1996); photograph courtesy of Laurie Crossey. (c) Four generations of ~1.69 Ga granite dikes in Spaghetti
Canyon of Lower Granite Gorge of Grand Canyon, note pen for scale. Earlier dikes underwent ductile deformation before later
dikes were emplaced, suggesting an interplay of diking and ductile flow at middle crustal depths (Karlstrom and Williams, 2006);
photograph courtesy of Laurie Crossey. See color plates section.
generate a thermal anomaly, and allow bodies extent do the contents of magma chambers
of magma to reside at higher melt fraction for represent fractionated melts extracted from the
longer periods of time than in enmasse sce- mantle or melts of the existing crust. Melting
narios (Petford and Gallagher, 2001). The flux of of the crust, and subsequent segregation and
magma in such settings becomes an important advection of this melt, can change the density
parameter. structure of the crust, but does not, in itself,
cause growth of the crust. Injection of mantle
2.2.1â•…Multiple intrusions and crustal melts, and fractionation to form low-density
melting efficiency melts and then rocks, however, can contrib-
One of the more interesting questions regard- ute mass to the crust over time (Grunder et al.,
ing the growth of continental crust is to what 2006). As shown in Figure 2.2, during single
intrusion scenarios when the intruding magma words, energy is not “wasted” by heating parts of
has the same composition as the country rock, the crust that never become molten. Conductive
little crustal melting can be expected. However, heat transfer is less efficient than this end mem-
real melting scenarios can involve different ber, however, as energy will diffuse over large
lithologies and multiple intrusive events. areas, and eventually radiate from the surface of
Regions of strong isotopic contrast, such as the Earth. Therefore, only a small fraction of the
old continental crust, can sometimes provide region that experiences temperature increases
a clear indication of crustal melting. However, will reach the crustal solidus. Using published
mafic portions of the crust, especially the lower results for a range of melt compositions, intru-
crust, can have little isotopic contrast rela- sion volumes, thermodynamic variables, and
tive to mantle melts, and major- and trace-ele- melt fraction relationships, we can compare
ment consideration of the magmas can lead to the amount of crustal melt predicted by these
ambiguous interpretations of whether a magma models (VCmod) with the completely efficient end
is primarily of mantle or crustal origin (Hart et member:
al., 2002). Also, cryptic assimilation, or assimila-
VCmod
tion of crystal phases without completely melt- E% = 100 . (2.12)
ing them, can modify whole rock isotopic values VCeff
and generate mixed signatures even when the
full crustal melting reaction has not proceeded A 100% efficient scenario would convert all the
to high melt fraction (Wolff et al., 2002; Beard enthalpy from an intrusion into melting the
et al., 2005). A complementary tool to geochem- crust. A compilation of several models is shown
ical investigation of fractionation and melting in Table 2.1.
involves thermal calculations (e.g., Fig. 2.2). An almost universal conclusion from ther-
Many thermal models have been considered mal models, especially those considering multi-
for different melting scenarios, including ana- dimensional thermal diffusion, is that injection
lytical solutions applied to simplified scenarios, of mantle melts in the crust, even under optimal
parametric and numerical convection simula- conditions, is a relatively inefficient method of
tions, and numerical conduction simulations melting the crust. Although there are obvious
(Younker and Vogel, 1976; Wells, 1980; Huppert differences depending on the intruding magma
and Sparks, 1988; Bergantz, 1989; Pedersen et al., and country-rock compositions considered,
1998; Bittner and Schmeling, 1995; Barboza and most are only around 10–40% efficient. This
Bergantz, 1996; Raia and Spera, 1997; Petford means that even in relatively fertile, hydrous
and Gallagher, 2001; Annen and Sparks, 2002; lithologies, large volumes of intruded mantle
Dufek and Bergantz, 2005a). One way that these basalts are required to produce voluminous
models can be compared is in their relative effi- crustal melts.
ciency in melting the crust. An enthalpy balance Making the space required to accommodate
of a volume of basalt (VB) intruded at its liquidus intrusions has long been recognized as a con-
(TliqB) and cooling to a temperature T gives the straint on magma system growth, especially
volume of crust melted: where extensive fractionation is required, and
is sometimes termed the “room problem”.
VCeff =
(
ρB VB c pB (TliqB − T ) + LB )
. (2.11)
While crustal melting has been invoked as a
(
ρC c pC (TsolC − TR ) + ρC c pC (T − TsolC ) + LC f ) mechanism to alleviate the room problem and
explain the absence of observed fractionate
Here the B and C subscripts refer to basalt and residua, it can be as inefficient as fractionation
crust, respectively, and TR is the initial rock tem- and sometimes more inefficient than fraction-
perature. This volume can be considered the ation at producing silicic melts. (Fractionation
perfectly efficient end member, as all the energy of a basalt to high silicate rhyolite requires
from the intrusion goes into heating only the ~90% removal of crystals; Grout, 1926; Winter,
portion of crust that becomes molten. In other 2001.)
The crustal lithology, thickness, melt flux, coupled to thermal, chemical and rheological
and evolving thermal anomaly all play a role processes within and around the chamber.
in the synoptic view of crustal melting and Magma chamber stresses are generated
fractionated mantle melts. When considering internally by the pressurization or buoyancy of
the random intrusion of mafic melts in arc magma relative to lithostatic stresses, through
crustal conditions, Dufek and Bergantz (2005a) a variety of processes. Chamber over-pressuriza-
found that, in general, thin, immature arc crust tion, �P, can occur through injection of magma
undergoes relatively little melting, and that a into the chamber (Parfitt et al., 1993; Woods,
random sampling of melt from the crustal col- 1995), thermal expansion of magma in the
umn is likely to contain <â•›10% crustal melt by chamber or melting of country rocks (Bonafede,
volume. Thick and mature crust, after several 1990), and volume change upon crystallization
million years of intrusion for typical arc fluxes, of mineral phases (Fowler and Spera, 2008) fol-
results in greater crustal melting: with a crust lowing the thermodynamic relationship
of ~50 km and fertile amphibolite lower crustal
1 ∆V α
lithology, they found a random sampling could ∆P = + ∆T , (2.13)
result in roughly equal probabilities of intersect- β V β
ing crustal melts or fractionated mantle melts.
Due to the relative inefficiency of fraction- where ∆V and ∆T are changes in the chamber
ation and crustal melting to produce silicic melts volume and temperature. Both the magma com-
(the former requires large volumes of basalt as a pressibility β (~10–11 Pa–1) and thermal expansiv-
direct mass source and the latter requires large ity α (~10–5 K–1) depend on the major element
volumes to account for the required enthalpy composition and volatile content of the magma,
source), the accommodation mechanism and tec- as well as depth (Dobran, 2001; Rivalta and
tonic regime of the crust are also likely import- Segall, 2008). While the volume change during
ant factors in the production of evolved melts in phase change varies with mineral species and
the crust. The thermal and mass requirements magma composition, the net effect is a pressure
for magma chamber development and preserva- increase during melting and a pressure decrease
tion are inextricably linked to the tectonic and during solidification (Lange and Carmichael,
chamber-induced stress regimes that permit 1990; Rushmer, 1995; Dobran, 2001). Melt buoy-
accommodation. ancy results from progressive crystallization of
mineral phases from the melt (Marsh, 1996),
which can progressively increase the density
2.3╇ ╇Crustal stresses and magma difference ∆ρ between magma and host rock.
chambers Magma buoyancy may be important for long-
distance transport in dikes (Rubin, 1995; Roper
Stresses around magma chambers provide and Lister, 2005; see Chapter 3), and for cal-
the mechanical means by which transport of dera-forming eruptions (Gudmundsson, 1998;
magma to and from the chamber occurs, and McLeod, 1999; Pinel and Jaupart, 2005). Stresses
determine the nature of deformation and failure may also accumulate as volatile species (primar-
of surrounding country rocks. Sufficiently high ily H2O and CO2) exsolve from the melt (Sparks
stresses above lithostatic confinement (devia- et al., 1977; Tait et al., 1989). Volatile exsolution
toric stresses) will induce fracture and chamber may also lead to anisotropic stresses as bubbles
rupture (Tait et al., 1989), while isotropic mag- (super-critical fluid at these depths) rise toward
matic stresses such as uniform loading may the chamber roof (Woods and Cardoso, 1997).
variously cause or suppress chamber rupture The build-up of internal stress due to magma
(Vigneresse and Tikoff, 1999). Chamber stresses addition or withdrawal, volatile exsolution, and
are also responsible for the dynamic evolution crystallization depends on the chamber depth
of magma chambers in the broader context of and solubility relationships but generally can be
the entire volcanic system, and are strongly described by (Woods and Huppert, 2003)
exterior stress trajectories (Meriaux and Lister, stresses σ for a pressurized and buoyant fluid-
2002), and may be focused towards it as they filled sphere of radius (Rd) are, in a spherical
rise from greater depths (Karlstrom et al., 2009). polar coordinate system (r,ϕ,ϑ) (Fung, 1965),
Modeling stresses around magma chambers
provides a means of predicting where cham- −∆PRd3 (υ + 1)
dr =
ber rupture, and hence volcanic eruption, is 2Er 2
likely to occur (Muller and Pollard, 1977; Pinel 2∆ ρ gRd3 υ2 − 1 Rd2 (υ + 1)(4υ − 3
+ − cos ϕ,
and Jaupart, 2005). Conversely, knowledge of E (13υ − 11) r 2r 3
chamber-induced stresses at depth provides a
means to estimate the spatial extent of a par-
∆ ρ gRd3 (r 2 − Rd2 )(υ + 1)(4 υ − 3)
ticular volcanic system (Muller et al., 2001; dϕ = sin ϕ, (2.19)
2E (13υ − 11) r 3
Karlstrom et al., 2009). The influence of cham-
ber stresses on dike generation and focusing
dϑ = 0,
at depth thus has important consequences for
magma chamber stability in the volcanic system
as a whole. and
To illustrate the role of magma cham-
ber stress we consider some typical idealized ∆PRd3 ∆ ρ gRd3 (r 2 (υ − 2) + 3 Rd2 (4 υ − 3)) cos ϕ
σ rr = + ,
models. Because of the large viscosity contrast r3 r 4 (13υ − 11)
between magma and rocks at mid to shallow
crustal depths, magma chambers are commonly −∆PRd3
σ ϕ ϕ =σ ϑ ϑ =
modeled as cavities in an elastic medium, solv- 2r 3
ing the equations of Linear Elasticity (Fung, ∆ ρ gRd3 (r 2 (υ − 2) + 3 Rd2 (4 υ − 3) cos ϕ
− ,
1965): 2r 4 (13υ − 11)
∆ ρ gRd3 ( Rd2 − r 2 )(4υ − 3) sin ϕ
σ rϕ = ,
∂ 2d i 1 ∂ ∂d j 2r 4 (13υ − 11)
+ = f bi , (2.18)
∂x 2j 1 − 2υ ∂x i ∂x j
(2.20)
where d is the displacement vector, fb is the
body force vector, and υ is Poisson’s ratio. σ ρϑ = σ ϕ ϑ =
Application of Eq. (2.18) commonly assumes
that no body forces exist in the elastic medium, where ∆P is magma overpressure relative to
which neglects free surface effects. Most studies lithostatic pressure, and ∆ρg is the magma
consider static stresses only, with some excep- buoyancy. For isotropic overpressure alone, nor-
tions (Dragoni and Magnanensi, 1989; Bonafede, mal stresses decrease away from the chamber as
1990). ∆P/r3, while for buoyancy alone, stresses decrease
Many three-dimensional boundary value as ∆ρg/r2.
problems in elasticity are far from trivial, so A spherical chamber in an infinite elastic
studies of magma chamber stresses, especially medium approximates the qualitative aspects of
those that attempt to couple pure elasticity to static stresses, especially in the far field (Sartoris
heat transfer or fluid dynamics, tend to use very et al., 1990), but does not resolve higher-order
simplified chamber geometries and boundary moments if the surface boundary conditions
conditions. It is of note, however, that a rich become important. In particular, deviations
library of mathematical techniques for solving from spherical geometry result in concentra-
elasticity problems has been developed (Fung, tion of stresses in regions of high curvature
1965). around the chamber (Dieterich and Decker,
The simplest and most commonly used model 1975; Gudmundsson, 2006). Eshelby (1957) pro-
for a magma chamber is a spherical inclusion in vides an analytic solution for stresses around a
an infinite elastic medium. Displacements d and pressurized ellipsoidal inclusion in an infinite
. (2.21)
principal stress just outside the wall of a pressurized oblate ∆P (1 − υ) Rd3 r
spheroidal cavity in an infinite elastic medium, normalized by d r ( r , z = 0) =
µ ( r 2 + l 2 )3 / 2
the deviatoric stress of a spherical chamber. Stresses plotted
are compressive, and are oriented perpendicular to the
This is equivalent to the first term in a power
wall of the chamber. Progressive flattening of the spheroid
series expansion of stress boundary conditions
(increasing the aspect ratio a/c) results in the concentration
of deviatoric stresses in areas of high curvature. In the around a pressurized spherical cavity, match-
limit that a/c → ∞, the region near the medial plane of the ing boundary conditions on the sphere and the
ellipsoid develops a stress singularity. free surface in alternating, higher-order terms
(McTigue, 1987). Two-dimensional solutions to
this problem are algebraically more simple, and
elastic medium that serves as a good example of are given in the bipolar coordinate system by
such stress concentration. Figure 2.4 illustrates Jeffery (1921).
the stress concentration due to curvature by Other boundary conditions are also of inter-
taking the particular case of an oblate spheroid est in a careful treatment of magma chamber
(a = b > c), an appropriate approximate geometry stresses. Depth-dependent density is neglected
for a sill (Fialko et al., 2001), and gradually flat- in most elastic treatments of chamber stresses,
tening it (decreasing c). It is evident that devia- yet this can be important (Grosfils, 2007).
toric stresses are progressively concentrated Rheological change due to prolonged heating
toward the medial plane. Indeed, the limit c → 0 and partial melting of country rocks around
is a “penny-shaped” crack solution (Jaeger et al., a chamber is expected to produce a “shell” of
2007), for which it can be shown that there is material that behaves viscoelastically, and the
a 1/√r divergence of the normal stress at the presence of this material strongly affects cham-
edge of the ellipsoid. Other, more complicated ber stresses (Bonafede et al., 1986; Dragoni
chamber geometries have also been studied and Magnanensi, 1989; Jellinek and DePaolo,
with numerical methods (Dieterich and Decker, 2003; Newman et al., 2006b). In particular,
1975; Grosfils, 2007), but the qualitative features the presence of a viscoelastic shell provides a
of solutions are the same. mechanism by which high magma chamber
Another case of interest to the evolution of overpressures may be accommodated with-
the volcanic system occurs when the Earth’s out large-scale chamber failure, and thus may
surface, a stress-free boundary, begins to affect explain how large, high melt-fraction chambers
concentration of stresses around magma can grow in the first place (Jellinek and DePaolo,
chambers, such as during caldera formation 2003). Thermoelastic stresses due to the heating
(Gudmundsson, 1998). This occurs for shallow or cooling of magma have also been addressed,
and exert stresses comparable to other sources strength beyond the percolation threshold, i.e.,
(e.g., Bonafede 1990). when crystals form a rigid connected framework
The fact that such varied boundary condi- (Vigneresse et al., 1996; Saar and Manga, 2002;
tions affect the first-order mechanical evolution Caricchi et al., 2007; Champallier et al., 2008).
of magma chambers, both in theory and observa- Magmas also contain dissolved, and com-
tion, is highly suggestive that chamber stresses monly exsolved, volatile phases, mostly
are strongly coupled to the dynamic thermal H2O, CO2, and SO2. The fractions and relative
and chemical evolution inside and around the importance of these phases depend on the
chamber (Scandone et al., 2007). From a model- tectonic setting from which the magma origi-
ing perspective this requires coupling elastic, nates. As a consequence of crystallization and
thermal, and fluid dynamic equations, and thus cooling of the magma, saturation conditions
often requires numerical simulation (Gerya et al., change and the exsolved volatile volume frac-
2004). tions increase. This additional phase provides
a new source of buoyancy and affects both the
flow dynamics and the rheology of the mixture
(Eichelberger, 1980; Huppert et al., 1982; Rust
2.4╇ ╇ Magma chamber convection and Manga, 2002; Longo et al., 2006; Ruprecht
et al., 2008).
Convective motion in magma chambers helps When a magma with different physical prop-
determine the rate of magma differentiation, erties is injected into the chamber, or when the
sets the length scale of compositional heteroge- residing magma undergoes differentiation by
neities in these bodies, and redistributes melts convective fractionation (Sparks and Huppert,
with different physical properties, and thus may 1984), this may induce large variations in trans-
be responsible for the triggering and stalling port coefficients (e.g., viscosity) and density. The
of eruptions (Sparks et al., 1977; Snyder, 2000). amount of mixing or segregation between dif-
Injection and storage of magma at upper or ferent magmas depends on the injection rate,
middle crustal depth can lead to a large ther- the buoyancy of the injected magma, and the
mal disequilibrium between the host rocks and viscosity ratio (Turner and Campbell, 1986;
the magma body. This thermal disequilibrium Koyaguchi and Blake, 1991; Jellinek and Kerr,
drives buoyant motion through crystallization, 1999). The presence of multiple sources of buoy-
gas exsolution, or thermal expansion of the ancy, such as temperature and compositional
magma. differences between the magma batches, can
As magmatic systems cool below their liq- lead to a range of fluid dynamical instabilities
uidus temperatures, denser crystals generally (Chen and Turner, 1980; Sparks and Huppert,
form at the cooling boundaries and can be car- 1984; Turner and Campbell, 1986).
ried by gravitational instabilities from the roof
into the bulk of the convecting body where they 2.4.1â•… Rayleigh–Bénard convection
will remain in suspension or settle downwards Before addressing the complex dynamics of
(Sparks and Huppert, 1984; Martin and Nokes, magma bodies or even their simpler labora-
1989; Bergantz and Ni, 1999). The presence tory analogs, we review Rayleigh–Bénard
of this secondary phase can have an import- convection (Bénard, 1900; Rayleigh, 1916).
ant effect on the convection style, even at low Rayleigh–Bénard convection occurs in a fluid
crystal fraction (<â•›10 vol. %; Martin and Nokes, of constant thermal properties heated from the
1989; Koyaguchi et al., 1990, 1993; Sparks et al., base and cooled from above. The set of equa-
1993). At higher crystal fractions (closer to the tions describing the problem consists of mass,
solidus), long-range (1/r) interactions between momentum, and energy conservation. We use
crystals will start to play a dominant role in the Boussinesq approximation, where the dens-
the rheology of the mixture, as demonstrated ity of the fluid is assumed to vary linearly with
by hindered settling and even the onset of yield temperature for the buoyancy force term and is
(below a few volume percent) where the com- alternative process leads to (limited) mixing
plexity introduced by crystal–crystal interac- (Eichelberger, 1980; Bergantz and Breidenthal,
tions can typically be neglected. Most crystalline 2001).
phases follow convective motion of the magma One approach in analyzing the motion of
closely, except in very low viscosity magma disperse phases, such as crystals and bubbles in
(Martin and Nokes, 1989; Koyaguchi et al., 1993; magma, is to examine the path of individual crys-
Sparks et al., 1993). Some particles can, however, tals as well as the velocity field of the magma.
still be removed from the flow when magma stag- This approach, termed Eulerian–Lagrangian
nates and crystals reside for an extended period (here Eulerian refers to the continuous quan-
of time. Martin and Nokes (1989) described this tity, magma, and Lagrangian refers to the dis-
process with an exponential relationship assum- crete quantity, crystals) is usually favored for
ing that in the bulk of the chamber, the crystals dilute conditions when the coupling between
are well mixed and only in the bottom boundary the magma and crystal is essentially one-way
layer do crystals decouple from the flow. This (i.e., the magma imparts inertia and relative vel-
behavior also explains the results of Koyaguchi ocity to the crystals, but is not affected by drag
et al. (1993), who described periodicity between due to the crystals).
three regimes: (1) clarified chamber where crys- Lagrangian analysis is especially useful in
tals are deposited at the bottom, (2) well-defined the analysis of the trajectory of individual par-
layers (crystal-poor overlying crystal-rich), and ticles and can be used to better understand the
(3) wholesale overturn that mixes both layers. fluid volumes through which these particles
At shallow depths, water can exsolve from have passed. However, the large numbers of
the melt and form a dispersed bubble phase. individual particles in magmatic systems makes
The exsolution of volatiles strongly affects the this approach computationally unattractive.
fluid dynamics of magma chambers, generally An alternative that still preserves the ability to
resulting in large buoyancy contrasts within the compute phase relative motion is to treat the
melt (Huppert et al., 1982; Phillips and Woods, discrete phase as an effective continuum, or
2002). The role of exsolved volatiles has been Eulerian phase. In this approach, bubbles or
investigated for triggering eruptions (Snyder, crystals of the same diameter and density are
2000; Huppert and Woods, 2002; Fowler and averaged into a fluid and are represented locally
Spera, 2008) and for increasing the buoyancy of by a volume fraction. In the averaging process,
an unsegregated melt–bubble mixture (Bergantz details of particle trajectories and histories are
and Breidenthal, 2001; Ruprecht et al., 2008). lost, but two-way interaction between the par-
A particular example of some of these effects ticles and magma can be computed for realistic
is magma chamber recharge, where a mafic volume fractions of discrete particles.
magma is injected at the base of a chamber This approach is often referred to as Eulerian–
filled with a more silicic magma (Eichelberger, Eulerian, multi-fluid, or multi-continuum. Each
1980; Huppert et al., 1982). As the mafic intru- continuum (magma, crystals, bubbles) is rep-
sion cools, it partially crystallizes and exsolves resented by a conservation equation for mass,
bubbles. If enough bubbles exsolve and remain momentum, and thermal energy, along with
in suspension, this situation can lead to an constitutive relations that describe the density,
unstable density stratification where the under- rheology, and thermal properties of these con-
plating magma becomes lighter than the host, tinua. The set of equations is coupled through
causing overturn of the layer (Huppert et al., equal and opposite drag terms and terms for the
1982; Bergantz and Breidenthal, 2001; Ruprecht transfer of thermal energy.
et al., 2008). Conversely, if bubbles can separate We illustrate this approach with an example
from an underplating mafic melt, they will pond of a crystal-driven instability in a magma com-
at density interfaces to form an unstable foam posed of a melt and a denser crystal phase
layer from which volatile-rich plumes originate (Fig. 2.5). We assume that this basaltic cham-
and ascend through the less dense magma. This ber was intruded at its liquidus with a size
of 400 × 100 m, and has formed a 10 m thick steady-state convection, we relate the stirring
layer of crystals (10% crystals by volume) at the time to the strain rate
top of the chamber due to cooling. The crys-
1 ε H 2
tal-rich layer is denser than the magma (∆ρ = τm = log , (2.29)
300â•›kgâ•›m–3) and creates a density instability 2 ε D
with dripping crystal plumes (Bergantz, 1999).
Further information about the details of this where H2/D is the diffusion timescale of interest
approach, constitutive relations, and numerics in the mixing process. The dependence of the
can be found in Dufek and Bergantz (2005b, average strain rate on the Rayleigh number fol-
2007) and Ruprecht et al. (2008). lows (Coltice and Schmalzl, 2006):
We see from this example that density
κ
instabilities drive stirring in the entire chamber. ε = 0.023 Ra 0.685 . (2.30)
2
This example also illustrates the value of the H
multi-fluid approach in predicting the spatial and
temporal patterns of large-scale heterogeneity. Convection in magma chambers is, however,
strongly time-dependent, and as cooling pro-
2.4.3â•… Convection and mixing ceeds the mechanical properties of the mixture
Mixing is a result of the stretching and twist- vary strongly because of temperature changes
ing of flow paths by shear and normal strains and the presence of crystal phases. One meas-
integrated over time, and is important in deter- ure of cumulative mixing in unsteady convec-
mining compositional heterogeneity (or lack tion (Ra = Ra(t)) is the number of overturns
thereof ) in magma chambers. For simplicity experienced by the fluid (Huber et al., 2009).
we focus on mixing of a single-phase magma. The product of the stirring time (Eq. (2.29))
However, the concepts introduced here can and the strain-rate (Eq. 2.30) gives an estimate
be applied to more complex systems. Ottino of the total strain experienced by the fluid to
et al. (1979) formally described how fluid motion reach a point where molecular diffusion takes
at low Reynolds number leads to chaotic mix- over to further homogenize the system. This
ing. For large-wavelength heterogeneities, the quantity is largely invariant with respect to Ra
stretching induced by shear strain (ε̇̇) domi- (Figure 2.6(b), and therefore independent of
nates, and leads to a deformation proportional the dynamical history of the fluid (Huber et al.,
to ε̇̇–1 (Olson et al., 1984). Once the heterogenei- 2009). This offers a convenient alternative to
ties have been reduced to smaller sizes, normal quantify the mixing efficiency of a convective
strain accounts for the majority of the deform- fluid subjected to temporal changes. Newly
ation, which becomes proportional to exp(–2ε̇̇). created heterogeneities require about five
Following Coltice and Schmaltzl (2006), for overturns to be efficiently mixed.
Q â•›I magma mass flux into chamber ρc density of crystal phase (kg m–3)
(kg s–1) ρg density of gas phase (kg m–3)
Q â•›O magma mass flux out of chamber ρl density of liquid (melt) phase (kg m–3)
(kg s–1) ρB basalt density (kg m–3)
r, ϕ, ϑ spherical polar coordinates ρC crust density (kg m–3)
R* dimensionless rate of melt production σij stress tensor (Pa)
Rd magma chamber radius (m) σd deviatoric stress (Pa)
t time (s) σext external (far-field) stress (Pa)
T temperature (K) σint internal stress (Pa)
�V change in temperature (K) σlith lithostatic stress (Pa)
T0 reference temperature (K) σtot total stress in crust (Pa)
Tb temperature at magma chamber τm stirring time (s)
boundary (K) υ Poisson’s ratio
Tliq liquidus temperature (K) Bi Biot number, ratio of thermal
TliqB liquidus temperature of basalt (K) resistance of magma chamber versus
TR background temperature of country host rocks (= h/δk)
rock (K) Pe Peclet number, ratio of advective to
Tsol solidus temperature (K) diffusive heat transport (= cp ρ0δ u0/kḥ)
TsolC solidus temperature of crustal Pr Prandtl number, ratio of momentum
rock (K) diffusivity to thermal diffusivity
u0 reference velocity (m s–1) (= ν/κ)
V volume (m3) Ra Rayleigh numberPeclet, ratio of
�V volume change (m3) diffusion timescale to advection
V 0 total volume of the magma chamber timescale (= αg∆TH3/νκ)
(melt + crystals + bubbles; m3) Ste Stefan number, ratio of sensible to
V g volume of gas phase (m3) latent heat (= cpT0/L)
Vl volume of liquid (melt) phase (m3)
VCeff volume of crust melted at 100%
�efficient enthalpy transfer (m3) Acknowledgments
VCmod volume of crustal melt predicted
in thermal model (m3) This work was supported in part by NASA
u velocity field (m s–1) 08-MFRP08–0073, NSF 0948532, and a Swiss
x, x, y, z Cartesian coordinates (m) Postdoctoral Fund Fellowship.
α thermal expansion coefficient (K–1)
β compressibility (Pa)
δ thermal boundary layer thickness (m)
δiz Kronecker-delta (= 1 when i = z and References
zero when i ≠ z)
Ajakaiye, D. E. (1970). Gravity measurements over
ε̇̇ strain rate (s̃δ)
the Nigerian Younger granite province. Nature,
ϕv volume fraction of volatiles
225, 50–52.
η dynamic viscosity (Pa s) Anderson, E. M. (1936). Dynamics of the
κ thermal diffusivity (m2 s–1) formation of cone-sheets, ring-dykes, and
κeff effective thermal diffusivity (m2 s–1) cauldron-subsidences, Proceedings of the Royal Society
μ modulus of rigidity of wall rock (Pa) of Edinburgh, 56, 128–157.
ν kinematic viscosity (= η/ρ; m2 s–1) Anderson, A. T. (1976). Magma mixing€– Petrological
ρ magma density (kg m–3) process and volcanological tool. Journal of
�ρ density difference between magma Volcanology and Geothermal Research, 1(1), 3–33.
and country rock (kg m–3) Annen, C. and Sparks, R. S. J. (2002). Effects of
ρ0 reference density (kg m–3) repetitive emplacement of basaltic intrusions on
the thermal evolution and melt generation in the Bittner, D. and Schmeling, H. (1995). Numerical
crust. Earth and Planetary Science Letters, 203(3–4), modelling of melting processes and induced
937–955. diapirisms in the lower crust. Geophysical Journal
Bachmann, O., Miller, C. F. and de Silva, S. L. (2007a). International, 123(1), 59–70.
The volcanic-plutonic connection as a stage for Blundy, J. and Wood, B. (2003). Mineral-melt
understanding crustal magmatism. Journal of partitioning of uranium, thorium and their
Volcanology and Geothermal Research, 167, 1–23. daughters. In Uranium-Series Geochemistry, ed. B.
Bachmann, O., Oberli, F., Dungan, M. A. et al. (2007b). Bourdon, G. M. Henderson, C. C. Lundstrom and P.
40Ar/39Ar and U-Pb dating of the Fish Canyon S. Turner. Washington, DC: Minerological Society
magmatic system, San Juan Volcanic Field, of America and Geochemical Society, pp. 59–123.
Colorado: Evidence for an extended crystallization Bonafede, M. (1990). Axi-symmetric deformation
history. Chemical Geology, 236, 134–166. of a thermo-poro-elastic half-space: inflation of a
Barboza, S. A. and Bergantz, G. W. (1996). Dynamic magma chamber. Geophysical Journal International,
model of dehydration melting motivated 183, 289–299.
by a natural analogue: applications to the Bonafede, M., Dragoni, M. and Quareni, F. (1986).
Ivrea-Verbano zone, northern Italy. Transactions of Displacement and stress produced by a center of
the Royal Society of Edinburgh, 87, 23–31. dilation and by a pressure source in a viscoelastic
Beard, J. S., Ragland, P. C. and Crawford, M. L. half-space: application to the study of ground
(2005). Reactive bulk assimilation: A model for deformation and seismicity at Campi Flegeri,
crust-mantle mixing in silicic magmas. Geology, Italy. Geophysical Journal of the Royal Astronomical
33(8), 681–684. Society, 87(2), 455–485.
Bénard, H. (1900). Les tourbillons cellulaires dans Burgisser, A., Bergantz, G. W. and Breidenthal, R. E.
une nappe liquide. Revue générale des sciences pures (2005). Addressing complexity in laboratory
et appliquées., 11, 1261–1271. experiments: the scaling of dilute multiphase
Bergantz, G. W. (1989). Underplating and partial flows in magmatic systems. Journal of Volcanology
melting: Implications for melt generation and and Geothermal Research, 141(3–4), 245–265.
extraction. Science, 245, 1093–1095. Caricchi, L., Burlini, L., Ulmer, P. et al. (2007).
Bergantz, G. W. (1990). Melt fraction diagrams: the Non-Newtonian rheology of crystal-bearing
link between chemical and transport models. In magmas and implications for magma ascent
Modern Methods of Igneous Petrology: Understanding dynamics. Earth and Planetary Science Letters, 264,
Magmatic Processes, ed. J. Nicholls and J. K. Russell. 402–419.
Reviews in Mineralogy, Mineralogical Society of Carrigan, C. R. (1988). Biot number and thermos
America, pp. 240–257. bottle effect: Implications for magma-chamber
Bergantz, G. W. (1995). Changing paradigms and convection. Geology, 16, 771–774.
techniques for the evaluation of magmatic Castro, A. and de la Rosa, J. (1994). Nomarski study
processes. Journal of Geophysical Research, 100(B9), of zoned plagioclases from granitoids of the
17 603–17 613. Seville Range batholith, SW Spain, Petrogenetic
Bergantz, G. W. and Breidenthal, R. E. (2001). implications. European Journal of Mineralogy, 6,
Non-stationary entrainment and tunneling 647–656.
eruptions: A dynamic link between eruption Champallier, R., Bystricky, M. and Arbaret, L. (2008).
processes and magma mixing. Geophysical Research Experimental investigation of magma rheology
Letters, 28(16), 3075–3078. at 300 MPa: From pure hydrous melt to 76 vol %
Bergantz, G. W. and Ni, J. (1999). A numerical study of crystals. Earth and Planetary Science Letters, 267,
of sedimentation by dripping instabilities in 571–583.
viscous fluids. International Journal of Multiphase Chapman, D. S. (1986). Thermal gradients in
Flow, 25, 307–320. the continental crust. Geological Society Special
Berlo, K., Turner, S., Blundy, J., Black, S. and Publication, 24, 63–70.
Hawkeswork, C. (2006). Tracing pre-eruptive Chapman, D. S. and Furlong, K. P. (1992). Thermal
magma degassing using (210Pb/226Ra) disequilibria state of the continental lower crust. In
in the volcanic deposits of the 1980–1986 eruption Continental Lower Crust, ed. D. M. Fountain,
of Mount St. Helens Earth and Planetary Science R. Arculus, and R. W. Kay. Amsterdam: Elsevier,
Letters, 249(3–4), 337–349. pp. 179–198.
the interpolation and extrapolation of liquid-solid of Central Chile. Contributions to Mineralogy and
equilibria in magmatic systems at elevated Petrology, 98, 455–489.
temperatures and pressures. Contributions to Hopson, C. A. and Mattinson, J. M. (1994). Chelan
Mineralogy and Petrology, 119, 197–212. Migmatite Complex, Washington: Field evidence
Green, D. H. (1973). Experimental melting for mafic magmatism, crustal anatexis, mixing
studies on a model upper mantle composition and protodiapiric emplacement. In Geologic Field
at high-pressure under water-saturated and Trips in the Pacific Northwest: 1994 Geological Society of
water-undersaturated conditions. Earth and America Annual Meeting, ed. D. A. Swanson and R.
Planetary Science Letters, 19(1), 37–53. A. Haugerud. Seattle: Department of Geological
Grosfils, E. B. (2007). Magma reservoir failure on Sciences, University of Washington, pp. 1–21.
the terrestial planets: Assessing the importance Huber, C., Bachmann, O. and Manga, M. (2009).
of gravitational loading in simple elastic models. Homogenization processes in silicic magma
Journal of Volcanology and Geothermal Research, 166, chambers by stirring and latent heat buffering.
47–75. Earth and Planetary Science Letters, 283, 38–47.
Grout, F. F. (1918). Two phase convection in igneous Huppert, H. E. and Sparks, R. S. J. (1988). The
magmas. Journal of Geology, 26, 481–499. generation of granitic magmas by intrusion of
Grout, F. F. (1926). The use of calculations in basalt into continental crust. Journal of Petrology,
petrology€– A study for students. Journal of Geology, 29(3), 599–624.
34(6), 512–558. Huppert, H. E. and Woods, A. W. (2002). The role of
Grunder, A. L., Klemetti, E. W., Feeley, T. C. and volatiles in magma chamber dynamics. Nature,
McKee, C. M. (2006). Eleven million years of arc 420, 493–495.
volcanism at the Aucanquilcha Volcanic Cluster, Huppert, H. E., Sparks, R. S. J. and Turner, J. S. (1982).
Northern Chilean Andes: implications for the life Effects of volatiles on mixing in calc-alkaline
span and emplacement of plutons. Transactions of magma systems. Nature, 297, 554–557.
the Royal Society of Edinburgh, 97, 415–436. Hutton, J. (1788). Theory of the Earth; or an
Gudmundsson, A. (1998). Formation and investigation of the laws observable in the
development of normal-fault calderas and the composition, dissolution and restoration of the
initiation of large explosive eruptions. Bulletin of land upon the globe. Transactions of the Royal Society
Volcanology, 60, 160–170. of Edinburgh, 1, 209–304.
Gudmundsson, A. (2006). How local stresses control Ilg, B., Karlstrom, K. E., Hawkins, D. and
magma-chamber ruptures, dyke injections, and Williams, M. L. (1996). Tectonic evolution of
eruptions in composite volcanoes. Bulletin of paleoproterozoic rocks in Grand Canyon, Insights
Volcanology, 79, 1–31. into middle crustal processes. Geological Society of
Gudmundsson, M. T. and Hognadottir, T. (2007). America Bulletin, 108, 1149–1166.
Volcanic systems and calderas in the Vatnajokull Jaeger, J., Cook, N. G. and Zimmerman, R.
region, central Iceland: Constraints on crustal (2007). Fundamentals of Rock Mechanics, 4th edn.,
structure from gravity data. Journal of Geodynamics, Wiley-Blackwell.
43(1), 153–169. Jeffery, G. B. (1921). Plane stress and plane strain in
Hart, G. L., Johnson, C. M., Shirey, S. B. and Clynne, bipolar co-ordinates. Philosophical Transactions of the
M. A. (2002). Osmium isotope constraints on Royal Society of London A, 221, 265–293.
lower crustal recycling and pluton preservation Jellinek, A. M. and DePaolo, D. J. (2003). A model
at Lassen Volcanic Center, CA. Earth and Planetary for the origin of large silicic magma chambers:
Science Letters, 199, 269–285. precursors of caldera-forming eruptions. Bulletin of
Hibbard, M. J. (1981). The magma mixing origin of Volcanology, 65, 363–381.
mantled feldspars. Contributions to Mineralogy and Jellinek, A. M. and Kerr, R. C. (1999). Mixing and
Petrology, 76, 158–170. compositional stratification produced by natural
Hieronymus, C. F. and Bercovici, D. (1999). Discrete convection 2. Applications to the differentiation
alternating hotspot islands formed by interaction of basaltic and silicic magma chambers and
of magma transport and lithospheric flexure. komatiite lava flows. Journal of Geophysical Research,
Nature, 397(18), 604–607. 104(B4), 7203–7218.
Hildreth, W. and Moorbath, S. (1988). Crustal Jellinek, A. M., Manga, M. and Saar, M. O. (2004).
contributions to arc magmatism in the Andes Did melting glaciers cause volcanic eruptions in
eastern, California? Journal of Geophysical Research, CO2-rich magma. Geophysical Research Letters, 33,
109(B09206). doi:10.1029/2006GL027750.
Jessup, M. J., Jones, J. V., Karlstrom, K. E. et al. Manzella, A., Volpi, G. and Zaja, A. (2000).
(2006). Three Proterozoic orogenic episodes and New magnetotelluric soundings in the Mt.
an intervening exhumation event in the Black Somma-Vesuvius volcanic complex: preliminary
Canyon of the Gunnison region, Colorado. Journal results. Annali di Geofisica, 43(2), 259–270.
of Geology, 114, 555–576. Marsh, B. D. (1981). On the crystallinity,
Karlstrom, K. E. and Williams, M. L. (2006). Nature probability of occurrence, and rheology of lava
of the middle crust€– Heterogeneity of structure and magma. Contributions to Mineralogy and
and process due to pluton-enhanced tectonism: Petrology, 78, 85–98.
an example from Proterozoic rocks of the Marsh, B. D. (1989). Magma chambers. Annual Review
North American Southwest. In Evolution and of Earth and Planetary Sciences, 17, 439–474.
Differentiation of the Continental Crust, ed. M. Brown Marsh, B. D. (1996). Solidification fronts and
and T. Rushmer. Cambridge University Press, pp. magmatic evolution. Mineralogical Magazine, 60,
268–295. 5–40.
Karlstrom, K., Dufek, J. and Manga, M. (2009). Martin, D. and Nokes, R. (1989). A fluid-dynamical
Organization of volcanic plumbing through study of crystal settling in convecting magmas.
magmatic lensing by magma chambers and Journal of Petrology, 30, 1471–1500.
volcanic loads. Journal of Geophysical Research, 114, McLeod, P. (1999). The role of magma buoyancy in
B10204, doi:10.1029/2009JB006339. caldera-forming eruptions. Geophysical Research
Kavanagh, J. L., Menand, T. and Sparks, R. S. J. (2006). Letters, 26(15), 2299–2302.
An experimental investigation of sill formation McTigue, D. F. (1987). Elastic stress and deformation
and propagation in layered elastic media. Earth near a finite spherical magma body: resolution
and Planetary Science Letters, 245, 799–813. of the point-source paradox. Journal of Geophysical
Koyaguchi, T. and Blake, S. (1991). Origin of mafic Research, 92(B12), 12 931–12 940.
enclaves: Constraints on the magma mixing McTigue, D. F. and Mei, C. C. (1987). Gravity-induced
model from fluid dynamic experiments. In stresses near axisymmetric topography of
Enclaves and Granite Petrology. Amsterdam: small slope. International Journal of Numerical and
Elsevier. Analytical Methods in Geomechanics, 11, 257–268.
Koyaguchi, T. and Kaneko, K. (1999). A two-stage Meriaux, C. and Lister, J. R. (2002). Calculation of
thermal evolution model of magmas in dike trajectories from volcanic centers. Journal of
continental crust. Journal of Petrology, 40, 241–254. Geophysical Research, 107(B4), 2077–2087.
Koyaguchi, T., Hallworth, M. A., Huppert, H. H. and Mogi, K. (1958). Relationships between the eruptions
Sparks, R. S. J. (1990). Sedimentation of particles of various volcanoes and the deformation of
from a convecting fluid. Nature, 343, 447–450. the ground surface around them. Bulletin of the
Koyaguchi, T., Hallworth, M. A. and Huppert, H. E. Earthquake Research Institute, University of Tokyo, 36,
(1993). An experimental study on the effects of 99–134.
phenocrysts on convection in magmas. Journal of Muller, O. H. and Pollard, D. D. (1977). Stress state
Volcanology and Geothermal Research, 55, 15–32. near Spanish Peaks. Pure and Applied Geophysics,
Lange, R. L. and Carmichael, I. S. E. (1990). 115, 69–86.
Thermodynamic propreties of silicate liquids Muller, J. R., Ito, G. and Martel, S. J. (2001). Effects of
with emphasis on density, thermal expansion volcano loading on dike propogation in an elastic
and compressibility. In Modern Methods of Igneous half-space. Journal of Geophysical Research, 106(B6),
Petrology: Understanding Magmatic Processes, ed. J. 11101–11113.
Nicholls and J. K. Russell. Reviews in Mineralogy, Newman, A. V., Dixon, T. H. and Gourmelen,
Mineralogical Society of America, pp. 25–64. N. (2006a). A four-dimensional viscoelastic
Lees, J. M. (2007). Seismic tomography of magmatic deformation model for Long Valley Caldera,
systmes. Journal of Volcanology and Geothermal California, between 1995 and 2000. Journal of
Research, 167, 37–56. Volcanology and Geothermal Research, 150, 244–269.
Longo, A., Vassalli, M., Papale, P. and Barsanti, M. Newman, A. V., Dixon, T. H. and Gourmelen,
(2006). Numerical simulation of convection and N. (2006b). A four-dimensional visoelastic
mixing in magma chambers replenished with deformation model for Long Valley Caldera,
California, between 1995–2000. Journal of liquid fraction and composition, and source
Volcanology and Geothermal Research, 150, 244–269. permeability. Journal of Geophysical Research, 100, 15
Olson, P., Yuen, D. A. and Balsiger, D. (1984). 601–15 610.
Convective mixing and the fine structure of Rapp, R. P. and Watson, E. B. (1995). Dehydration
mantle heterogeneity. Physics of the Earth and melting of metabasalt at 8–32 kbar: Implications
Planetary Interiors, 36, 291–304. for continental growth and crust–mantle
Ottino, J. M., Ranz, W. E. and Macosko, W. (1979). recycling. Journal of Petrology, 36(4), 891–931.
A lamellar model for analysis of liquid-liquid Rayleigh, Lord (1916). On convection currents
mixing. Chemical Engineering Science, 34, 877–890. in a horizontal layer of fluid when the higher
Parfitt, E. A., Wilson, L. and Head, J. W. (1993). temperature is on the under side. Philosophical
Basaltic magma reservoirs: factors controlling Magazine, 32, 529–546.
their rupture characteristics and evolution. Reid, M. R. (2008). How long does it take to supersize
Journal of Volcanology and Geothermal Research, 55, an eruption? Elements, 4(1), 23–28.
1–14. Reid, M. R., Coath, C. D., Harrison, T. M. and
Patiño Douce, A. E. and Beard, J. (1995). McKeegan, K. D. (1997). Prolonged residence times
Dehydration-melting of biotite gneiss and quartz for the youngest rhyolites associated with the
amphibolite from 3 to 15 kbar. Journal of Petrology, Long Valley Caldera. Earth and Planetary Science
36, 707–738. Letters, 150, 27–39.
Patiño Douce, A. E. and Harris, N. (1998). Rivalta, E. and Segall, P. (2008). Magma
Experimental constraints on Himalayan anatexis. compressibility and the missing source for some
Journal of Petrology, 39, 689–710. dike intrusions. Geophysical Research Letters, 35,
Pedersen, T., Heeremens, M. and van der Beek, P. L04306, doi:10.1029/2007GL032521.
(1998). Models of crustal anatexis in volcanic Roper, S. M. and Lister, J. R. (2005). Buoyancy-driven
rifts: applications to southern Finland and the crack propagation from an over-pressurized
Oslo Graben, southeast Norway. Geophysical Journal source. Journal of Fluid Mechanics, 536, 79–98.
International, 132(2), 239–255. Rubin, A. M. (1995). Propogation of magma-filled
Petford, N. and Gallagher, K. (2001). Partial melting cracks. Annual Review of Earth and Planetary Sciences,
of mafic (amphibolitic) lower crust by periodic 23, 287–336.
influx of basaltic magma. Earth and Planetary Rudnick, R. L., McDonough, W. F. and O’Connell,
Science Letters, 193, 483–499. R. J. (1998). Thermal structure, thickness and
Phillips, J. C. and Woods, A. W. (2002). Suprresion composition of continental lithosphere. Chemical
of large-scale magma mixing by melt-volatile Geology, 145, 395–411.
separation. Earth and Planetary Science Letters, 204, Ruprecht, P., Bergantz, G. W. and Dufek, J.
47–60. (2008). Modeling of gas-driven magmatic
Pinel, R. and Jaupart, C. (2005). Caldera formation over-turn:Tracking of phenocryst dispersal
by magma withdrawal from a reservoir beneath a and gathering during magma mixing.
volcanic edifice. Earth and Planetary Science Letters, Geochemistry Geophysics Geosystems, 9,
230, 273–287. doi:10.1029/2008GC002022.
Poland, M., Hamburger, M. and Newman, A. V. Rushmer, T. (1991). Partial melting of two
(2006). The changing shape of active volcanoes: amphibolites: contrasting experimental results
History, evolution and future challenges for under fluid-absent conditions. Contributions to
volcano geodesy. Journal of Volcanology and Mineralogy and Petrology, 107, 41–59.
Geothermal Research, 150, 1–13. Rushmer, T. (1995). An experimental deformation
Prakash, C. (1990). Two-phase model for binary of partially molten amphibolite: Application to
solid-liquid phase change, part I: governing low-melt fraction segregation. Journal of Geophysical
equations. Numerical Heat Transfer, 18(B), 131–145. Research, 100, 15,681–15,695.
Raia, F. and Spera, F. J. (1997). Simulations of Rust, A. C. and Manga, M. (2002). The effects
crustal anatexis: Implications for the growth and of bubble deformation on the viscosity of
differentiation of continental crust. Journal of suspensions. Journal of non-Newtonian Fluid
Geophysical Research, 102(B10), 22 629–22 648. Mechanics, 104, 53–63.
Rapp, R. P. (1995). Amphibole-out phase boundary Saar, M. O. and Manga, M. (2002). Continuum
in partially melted metabasalt, its conrol over percolation for randomly oriented soft-core
prisms. Physical Review E, 65, doi:10.1103/ Touloukian, Y. S., Judd, W. R. and Roy, R.F. (ed.)
PhysRevE.65.056131. (1981). Physical Properties of Rocks and Minerals. New
Sartoris, G., Pozzi, J. P., Phillippe, C. and Mouel, J. L. York: McGraw-Hill.
L. (1990). Mechanical stability of shallow magma Tsuchida, E. and Nakahara, I. (1970).
chambers. Journal of Geophysical Research, 95(B4), Three-dimensional stress concentration around
5141–5151. a spherical cavity in a semi-infinite elastic body.
Scandone, R., Cashman, K. V. and Malone, S. D. Japan Society of Mechanical Engineering Bulletin, 13,
(2007). Magma supply, magma ascent and the style 499–508.
of volcanic eruptions. Earth and Planetary Science Turcotte, D. L. and Schubert, G. (1982). Geodynamics:
Letters, 253(3–4), 513–529. Applications of Continuum Physics to Geologic Problems.
Segall, P. and Davis, J. L. (1997). GPS applications New York: Wiley.
for geodynamics and earthquake studies. Annual Turner, J. S. and Campbell, I. H. (1986). Convection
Review of Earth and Planetary Sciences, 25, 301–336. and mixing in magma chambers. Earth Science
Simon, J. I. and Reid, M. R. (2005). The pace of Reviews, 23, 255–352.
rhyolite differentiation and storage in an Vazquez, J. A. (2004). Time scales of magma storage
‘archetypical’ silicic magma system, Long Valley, and differentiation beneath caldera volcanoes
California. Earth and Planetary Science Letters, from uranium-238-thorium-230 disequilibrium
235(1–2), 123–140. dating of zircon and allanite. PhD Thesis,
Snyder, D. (2000). Thermal effects of the intrusion University of California, Los Angeles.
of basaltic magma into more silicic magma Vigneresse, J. L. and Tikoff, B. (1999). Strain
chamber and implications for eruption partitioning during partial melting and
triggering. Earth and Planetary Science Letters, 175, crystallizing felsic magmas. Tectonophysics, 312,
257–273. 117–132.
Sparks, R. S. J. and Huppert, H. E. (1984). Density Vigneresse, J. L., Barbey, P. and Cuney, M. (1996).
changes during fractional crystallization of Rheological transitions during partial melting
basaltic magmas: Fluid dynamical implications. and crystallization with application to the felsic
Contributions to Mineraology and Petrology, 85, magma segregation and transfer. Journal of
300–309. Petrology, 37, 1579–1600.
Sparks, R. S. J., Sigurdsson, H. and Wilson, L. (1977). Voller, V. R. (1985). A heat balance integral method
Magma mixing: a mechanism for triggering acid for estimating practical solidification parameters.
explosive eruptions. Nature, 267, 315–318. IMA Journal of Applied Mathematics, 35, 223–232.
Sparks, R. S. J., Huppert, H. E., Koyaguchi, T. and Voller, V. R. and Cross, M. (1993). Solidification
Hallworth, M. A. (1993). Origin of modal and processes: Algorithms and codes. In Mathematical
rhythmic igneous layering by sedimentation in a Modelling for Materials Processing, ed. M. Cross, J. F.
convecting magma. Nature, 361, 246–249. T. Pittman and R. D. Wood. Oxford: Clarendon
Speer, J. A., Naeem, A. and Almohandis, A. A. (1989). Press.
Small-scale variations and subtle zoning in Voller, V. R. and Swaminathan, C. R. (1991). General
granitoid plutons: the Liberty Hill pluton, South source-based method for solidification phase
Carolina, U.S.A. Chemical Geology, 75, 153–181. change. Numerical Heat Transfer, 19B, 175–189.
Spera, F. J. and Crisp, J. A. (1981). Eruption volume, Wallace, G. and Bergantz, G. W. (2005). Reconciling
periodicity, and caldera area: Relationships and heterogeneity in crystal zoning data, an
inferences on developments of compositional application of shared characteristic diagrams at
zonation in silicic magma chambers. Journal of Chaos Crags, Lassen Volcanic Center, California.
Volcanology and Geothermal Research, 11, 169–187. Contributions to Mineralogy and Petrology, 149(1),
Spiegelman, M. and Kenyon, P. (1992). The 98–112.
requirements for chemical disequilibrium during Wells, P. R. A. (1980). Thermal models for
magma migration. Earth and Planetary Science the magmatic accretion and subsequent
Letters, 109, 611–620. metamorphism of continental crust. Earth and
Tait, S., Jaupart, C. and Vergniolle, S. (1989). Planetary Science Letters, 46, 253–265.
Pressure, gas content and eruption periodicity Winter, J. (2001). An Introduction to Igneous and
of a shallow, crystallizing magma chamber. Metamorphic Petrology. Upper Saddle River, NJ:
Geophysical Research Letters, 25(18), 3413–3416. Prentice-Hall.
Wolf, M. B. and Wyllie, P. J. (1994). the following quantities may be useful for your
Dehydration-melting of amphibolite at 10 kbar: Â�estimate: The latent heat of the magma is 2 × 105
the effects of temperature and time. Contributions Jâ•›kg–1, the density of the rock is 3000 kg m–3, heat
to Mineralogy and Petrology, 115, 369–383. capacity is 800â•›Jâ•›kg–1 K–1, and the conductivity is
Wolff, J. A., Balsley, S. D. and Gregory, R. T. (2002). 5.3 W m–1â•›K–1.
Oxygen isotope disequilibrium between quartz 2.2 Calculate the critical thickness of magma required
and sanidine from the Bandelier Tuff, New for convection. Assume a simplified magma
Mexico, consistent with a short residence time chamber with a large aspect ratio (thickness H ≪
of phenocrysts in rhyolitc magma. Journal of width W), and that the source of convection in the
Volcanology and Geothermal Research, 116(1–2), magma chamber is the density contrast between
119–135. crystal-rich plumes sinking from the roof and the
Woods, A. W. (1995). The dynamics of explosive ambient magma, estimated here to be ∆ρ = 100
volcanic eruptions. Reviews of Geophysics, 33(4), kg m3. Taking the bulk magma viscosity of (a)
495–539. basalt (η = 102 Pa s), (b) andesite (η = 104 Pa s),
Woods, A. W. and Cardoso, S. S. S. (1997). Triggering and (c) rhyolite (η = 105 Pa s), calculate the crit-
of basaltic volcanic eruptions by bubble-melt ical thickness of magma Hcr necessary for the
separation. Nature, 385, 518–520. onset of convection in the magma chamber,
Woods, A. W. and Huppert, H. E. (2003). On using
magma chamber evolution during slow effusive
eruptions. Journal of Geophysical Research, 108(B8), ∆ ρ gH cr
2403. Racr = = 103 ,
ηκ
Wyllie, P. J. (1977). Crustal anatexis: An
experimental review. Tectonophysics, 43, 41–71.
where ̇κ = 10–6 m2 s–1 is the thermal diffusivity.
Younker, L. W. and Vogel, T. A. (1976). Plutonism
Repeat the calculations for convection driven by
and plate tectonics: the origin of circum-Pacific
bubble plumes (density contrast ∆ρ = 500 kg m–3.
batholiths. Canadian Mineralogist, 14, 238–244.
Finally, compare these results with the case where
convection is driven by thermal expansion alone,
Δρ = ρ0αΔT, with ρ0 = 2500 kg m–3, α = 10–5 K–1 and
Exercises ∆T = 10 K.
2.1 A basaltic sill (1280â•›°C) intrudes a kilometer Online resources available at www.cambridge.
beneath a water-saturated sedimentary layer org/fagents
(where the surrounding rocks are ~300â•›°C). Using
simple scaling, estimate the amount of time that • Supplement 2A: Summary of dimensionless num-
will elapse before the layer begins to be heated bers relevant to volcanic processes
by the intrusion if heat transfer is only by con- • Additional exercises
duction and no phase change occurs. Some of • Answers to exercises
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
erupted at the surface or intruded as a hori- vertical motion, or extension and subsidence in
zontal or a vertical sheet. For this reason, the rift zones, for example, as responses to diking
emplacement of dikes has a fundamental influ- (Rubin, 1992; Buck, 2006). This class of models
ence on the thermal and geochemical structure pertains to the “before and after” of a given
and evolution of the lithosphere. All volcanic dike event, without modeling dike propagation
eruptions involve the process of dike propaga- in detail. Although this chapter does not discuss
tion as the means by which magma and asso- these models in detail, the two approaches are
ciated gases reach the surface. The motion of complementary.
dikes below the Earth’s surface can be detected
by dynamic geophysical signals such as seismic
swarms and real-time strain measurements.
Understanding the dynamics of fissure propa- 3.2╇ ╇ Field observations
gation is therefore of key importance for inter-
preting sequences of ground deformation and The most obvious geologic context in which to
seismic swarms prior to eruptions. Ultimately, think of dikes is that of rifting, whether along
a good understanding of such precursory sig- a continental rift, such as in East Africa, or at
nals is required during volcanic crisis man- ocean ridges. The significance of sheeted dike
agement. For example, advanced warning of complexes in obducted oceanic crust, consisting
the likely dynamics of an impending eruption, of essentially 100% dikes, has been understood
before a dike breaches the surface, would be for several decades. Nevertheless, dikes are not
highly beneficial. Dike propagation is a chal- limited to divergent plate boundaries€ – they
lenging physical problem, involving fluid flow, are the fundamental means by which magma
elasticity, fracturing, and heat transport. Strong is transferred to the surface during volcanic
rheologic gradients develop in the magma, activity, whatever the geodynamic context. The
which partially solidifies and degasses during relatively rapid propagation and flow through a
propagation. Deformation and fracture of the fissure is effective for preventing magma solidi-
host must be achieved in order for propagation fication during ascent. This section highlights
to proceed. This combination of physics sub- some recent episodes of dike propagation that
disciplines represents a complex mathematical have been captured in real time by geophys-
problem, but equally, a rich phenomenology ical instrument arrays, and which illustrate
can be anticipated. the dynamics of dike emplacement. To date,
This chapter focuses on the fluid modeling these recorded episodes are neither sufficiently
aspects of dike propagation, emphasizing the numerous nor accurately captured to have thor-
relationships between experiment and theory. oughly documented dike dynamics. Vertical
More detailed reviews of theoretical aspects propagation has been observed less frequently
of dike dynamics are given by Lister and Kerr than horizontal propagation, for example.
(1991) and Rubin (1995). When faulting and dike Furthermore, diking processes may vary among
injection take place simultaneously, the tectonic different tectonic stress regimes. Geologic evi-
strain is absorbed both by the emplacement of dence of swarms of mega-dikes (Halls and Fahrig,
magma and the faulting. Understanding this 1987) much larger than events observed by geo-
system therefore requires coupling of magma physical arrays, also suggests that gaps remain
fluid dynamics and fault mechanics. While such in the record of real-time observations. Before
studies are rare, there have been some inter- describing geophysically observed diking epi-
esting contributions at this research interface sodes, we briefly discuss insights into dike prop-
in recent years, notably building on intensive erties gleaned from geologic field observations.
studies of mid-ocean ridge and rifting processes.
Larger-scale models of rift zone evolution involv- 3.2.1â•… Dike geometry as seen in the field
ing repeated dike injection events are largely The geometry of dikes can be inspected dir-
concerned with explaining horizontal versus ectly in the field where they have been exposed
Figure 3.1╇╇ (a) Dike exposed by erosion on Piton des the dike at a given location. Hence there is no
Neiges, Reunion Island (photo courtesy of N.Villeneuve). guarantee that a solidified dike has the same
The quasi two-dimensional geometry of dikes commonly thickness as it did during emplacement. Despite
adopted in theoretical models can be appreciated in this view. this complication, field observations provide
Pervasive horizontal jointing was caused by lateral cooling. a first-order estimate of dike geometry. Silicic
(b) Dike emplaced just at the surface during the Nyiragongo dikes, in which the magma is typically several
eruption 2002 (photo courtesy of J.-C. Komorowski). A
orders of magnitude more viscous than basalt,
solidified crust can be seen but the magma that occupied
tend to be thicker than their basaltic counter-
the interior of the dike was drained during the eruption. See
color plates section.
parts by about an order of magnitude, although
examples of thin silicic dikes also exist. However,
in general there are fewer field studies of silicic
by erosion. Observations made in the Tertiary dikes than of basaltic dikes. In this chapter, we
Igneous Province in the UK (Jolly and Sanderson, do not emphasize magma composition as far as
1995) related to continental rifting, or in the older, liquid properties are concerned, although the
more eroded parts of Iceland (Gudmundsson, greater viscosity and volatile content are clearly
and Brenner, 2005) related to oceanic rifting, for important factors for dike emplacement and
example, indicate that basaltic dikes commonly eruption of silicic magmas.
have widths on the order of one to a few meters. While the basic picture of dike geometry is
The lengths and breadths (vertical and horizon- that of a two-dimensional sheet, it is clear that
tal extents, respectively) of dikes are variable, dike surfaces are rarely smooth. It is common
but always much greater than dike widths, and to observe damage in the host rocks in the form
thus it is natural to approximate dikes as two- of joints and cracks, and pieces of host rock
dimensional sheets Fig. 3.1(a). Nevertheless, spalling off the wall, as well as irregular offsets
geologic exposures that allow examination of of the dike profile when viewed in cross-section.
individual dikes over large vertical or horizontal Currently it is not clear whether such features
extents are rare, so it is difficult to characterize are important clues to propagation processes
dike geometry in greater detail. Moreover, when or simply secondary details, but they may be
a dike is propagating, pressure changes in the relevant in the context of characteristic seis-
magma and stress changes in the host will cause micity accompanying propagation. In notable
variations with time in the opening (width) of field studies on the Colorado Plateau, Delaney
and Pollard (1981) and Delaney et al. (1986) Iceland (e.g., Einarsson and Brandsdottir, 1980;
examined the damage recorded in host rocks, Björnsson, 1985; Buck et al., 2006; Pedersen et al.,
and attempted to quantify in situ resistance to 2007) and Hawaii (e.g., Pollard et al., 1983; Rubin
dike propagation. They concluded that a consid- and Pollard, 1987). Here, we highlight the gen-
erable amount of energy must be expended in eral features of such events that provide useful
irreversible damage to the host, and suggested constraints for propagation models, including
that effective fracture toughness might be a lot examples of rift-related, subduction-related, and
higher than values measured on small rock sam- hot-spot- (shield volcano-) related magmatism.
ples in the laboratory. The underlying quantita- In 2005 a dike ~60 km long and 5–8 m wide,
tive questions, related to use of Linear Elastic as constrained by satellite-borne radar interfer-
Fracture Mechanics (LEFM; see Section 3.3.2) in ometry, was injected along the Manda Harraro
theoretical models, concern the size of the zone segment of the rift in Afar, Ethiopia (Wright et al.,
of irreversible damage and stress levels therein 2006). Approximately a dozen earthquakes with
(e.g., Rubin, 1993a). magnitudes of 5 to 5.5 were recorded by regional
Field observations of chilled margins and networks over an interval of a few days. While
vesicles in dikes indicate that magma properties there was insufficient ground instrumentation
vary both spatially and temporally during dike close to the site of injection to observe accurately
propagation, although it is not straightforward the duration of emplacement, an upper bound
to relate the predictions of models of dynamic for the duration is ~1 week. A notable feature
processes to these static observations, which are of this event is the considerable length of the
“frozen in” by solidification of the magma. For first dike that was emplaced, which was some-
example, Figure 3.1(b) shows a dike just at the what larger than that of the eruptive fissure of
surface from which the magma drained laterally Laki 1783, and comparable with the length of
after formation of a chilled margin. In addition, typical mid-ocean ridge segments. The overall
heterogeneities in the host medium can affect volume of magma intruded in the September
dike propagation. Gudmundsson and Brenner 2005 Afar event was 1.5 km3 (Wright et al., 2006;
(2001) and Gudmundsson (2005) described the Grandin et al., 2009), approximately an order of
possible effects on dike trajectory through a geo- magnitude less than the volume erupted at Laki.
logic formation in which successive layers have In striking contrast to Laki, no basaltic magma
contrasting mechanical properties such as elas- reached the surface at Afar in September 2005.
tic stiffness, leading to stress refraction effects. A dozen or so smaller dikes have been injected
Kavanagh and Sparks (2011) emphasized the since the initial event (on average 1 dike every
importance of variable host rock properties in few months) from a central location on the rift
affecting the final geometry of a swarm of kim- segment (Ebinger et al., 2010). A small propor-
berlite dikes, and Baer and Hamiel (2010) inter- tion of these injections led to eruptions of bas-
preted a set of dikes in the Dead Sea region as altic magma, indicating that the volume of the
having propagated horizontally under the influ- dikes is in large part accommodating tectonic
ence of a density interface in the host rock. strain in the depth interval of 4–10 km, while
the strain at shallower depths is being taken up
3.2.2â•…Geophysically observed dike by motion on faults.
injection events At the time of writing, the Afar rifting epi-
In this section, we discuss episodes of dike injec- sode appears to be nearing its end after more
tion that have been recorded geophysically in than six years of activity. The lack of vegetation
real time, including the rifting event in Afar, in this region has allowed high€– quality radar
Ethiopia, which started in 2005 and is ongoing interferometry to constrain the volumes and
at the time of writing, the 1997–98 event near final positions of the dikes. Moreover, some of
the Izu Peninsula of Japan, and recent eruptions the later injections were recorded on more prox-
of Piton de la Fournaise. Valuable data have also imal, albeit sparse, seismic networks (Ebinger
been obtained for events elsewhere, notably in et al., 2008; Keir et al. 2009; Grandin et al., 2011).
Figure 3.2╇╇ Data obtained during dike emplacement in of crust at some mid-ocean ridges have identified
the central section of the Manda Harraro rift segment in magma storage/accumulation regions on the
Afar, Ethopia modified from Grandin et al. (2011). (a), (b) order of a kilometer in vertical and cross-ridge
Earthquake locations deduced from data acquired by a local extent (Singh et al., 2006). However, taking a
temporary seismic network; the white cross indicates the broader view of magmatic activity in continents
main source reservoir feeding the dikes. (c) Dike opening and oceans, observations that tightly constrain
constrained by satellite-borne radar interferometry (Grandin
magma distribution prior to dike emplacement
et al., 2009). Each contour represents an opening of 0.5 m.
are still quite rare. One challenge is to record a
The dashed line shows the position of the dike tip based on
a first-order interpretation that seismicity is centered on
sufficiently clear picture of dike dynamics that,
the propagating tip (Grandin et al., 2011). However, it is clear via a physical model, robust inferences can be
in (b) that seismicity at any given time is distributed over a made about tectonic stress boundary conditions
horizontal distance at least 5 km in width, i.e., considerably and the magma source pressure driving dike
greater than the size of a process zone that would be emplacement. For recent models that focus on
consistent with the LEFM framework. The majority of this how a sequence of dikes can be interpreted as
seismicity might therefore actually be located in front of or a response to a given set of conditions prior to
behind the tip.
magmatic activity, we refer the reader to Buck
(2006) and Buck et al. (2006), and here focus on
the dynamics of individual injections.
In Figure 3.2, seismic recordings of one such In Figure 3.2(b), one can envisage that the
injection clearly show the migration of events. dike tip is located at the barycenter of the
A general picture emerges of a shallow magma events and fit a curve to dike tip position ver-
source, positioned roughly centrally in the rift sus time, and hence deduce dike velocity by dif-
segment, from which magma migrates along ferentiation (Rivalta, 2010; Grandin et al., 2011).
the rift. This interpretation largely confirms The seismic events are sufficiently well located
earlier suggestions based on the pioneering that one can probably improve on this first-
use of microseismicity to follow dike injection order interpretation, but even if one locates the
during the Krafla rifting episode in the 1980s dike tip in this way, there is clearly seismicity
(Brandsdottir and Einarsson, 1979; Einarsson occurring at locations well behind and beyond
and Brandsdottir, 1980), and is similar to inter- the tip. It is reasonable to expect some seismi-
pretations of events that have been recorded city ahead of the dike tip, and to either side of
at submerged mid-ocean ridges (e.g., Blackman the dike close behind the tip, depending on the
et al., 2000; Tolstoy et al., 2006). Seismic images stress distribution (Rubin and Gillard, 1998).
(a) 0 (b)
km
0 1 2
B’
A
1 34°-58’
2
A’
B
3
139° 08’ 139° 10’ 139° 12’
4
(c) A A’ (d) B
Depth (km)
B’
5
4
6
Depth (km)
6
7
8 8
9
10
–3 –2 –1 0 1 2 3 –1 0 1
10 Distance (km)
0 1 2 3 4 5 6 7 8 9 10 11
Day from 20 April 1998 0 2 4 6 8 10 12 14 16 18 20
Day from 20 April 1998
(e)
2
Elevation (km)
inverted using a model of an opening and mov- injection shows intermittent rather than steady
ing dike by Aoki et al. (1999) for the 1997 seis- seismic emission as the dike moves towards the
mic swarm (Fig. 3.4) and by Okada et al. (2000) surface. These observations demonstrate the
for the 1998 seismic swarm. Morita et al. (2006) importance of transient processes in propaga-
combined seismic and geodetic data using the tion dynamics, which could arise due to host
former as a priori information while inverting rock heterogeneities (e.g., in elastic properties
the latter. These impressive studies start to pro- or density), or variations in magma properties
vide a picture of dike geometry during propaga- (e,g., due to cooling/�solidification or degassing).
tion. We conclude that propagation proceeded Sections 3.3.2 and 3.4.4 investigate the role of
in a somewhat discontinuous way with bursts of variable magma and host rock properties from
seismicity occurring around the margins of the theoretical and experimental perspectives, in
dike rather than a continuous migration of a an effort to aid in the interpretation of these
well-identified source. Seismic events took place field observations.
over a depth interval of 3–8 km throughout the This summary of three diking events that
duration of the swarm (Fig. 3.3), which supports were recorded by geophysical arrays sets the
the idea that earthquakes also occur well away scene for discussing modeling contributions
from the tip region. that have been made over the last 20 years or
Piton de la Fournaise, Réunion Island so. The observational data available at present
is a highly active volcano in which seismic are biased towards predominantly horizontal
swarms preceding eruptions provide an oppor- rather than vertical propagation, whereas much
tunity to follow dike motion (Battaglia et al., theoretical work has explicitly tackled vertical
2005; Lengliné et al., 2008). Other methods buoyancy-driven propagation. Section 3.3 sets
have recently been derived to analyze seismic out the basic mathematical description of a
swarms too dense to allow a classical approach propagating fissure, with an emphasis on lubri-
(Taisne et al., 2011a). Figure 3.3(e) shows an cation theory, initially for situations in which
example of migrating seismicity associated magma and host rock properties are homoge-
with a propagating dike recorded before it neous. We then broaden the discussion to more
reached the surface at Piton de la Fournaise. recent studies involving heterogeneous magma
Similarly to the relatively long-lived injection and/or host properties, which attempt to more
at Izu peninsula, this relatively short-lived closely model natural conditions.
∂h 1 ∂ 3 ∂Pe â•…
where V is the volume of magma in the
= h − ∆ρ g , (3.1a)
∂t 3 η ∂z ∂z fissure.
Figure 3.5 shows results for the shape of a
−G
zf â•…
∂h ∂Γ steadily propagating buoyant dike as a function
Pe = ∫
(1 − v )π −∞ ∂Γ Γ − z
, (3.1b)
of K*. To first order, the dike has a swollen head
region, where elastic pressure gradients are pre-
1− v 2 â•…
h (z, t ) ~ Kc (l − z ), (3.1c) sent, and the thickness of this head increases
G π with increasing K*. The head is followed by a
the fissure length follows l ~ t1/3, and the crack V = Atα, (3.5)
thins with time as it extends in length.
where A is a constant whose dimensions depend
The above theoretical work has inherent
on the value of the exponent α, and α = 1 for
simplifications when compared with natural
constant magma flux and α = 0 for constant
situations, perhaps the most obvious being
magma volume. Table 3.1 summarizes fissure
the assumptions of two-dimensionality and
propagation relationships for various model
of constant magma and host rock properties.
situations represented by balances between dif-
Before discussing more recent results that have
ferent driving and resistive forces. As a fissure
extended this basic theoretical approach to
lengthens, it can pass from a regime defined by
cases with variable magma or host properties,
the balance between elastic and viscous pres-
we introduce the approach of writing simple
sures (Pe vs. Pv), to one between buoyancy and
physical balances to obtain scaling laws. This
viscous pressures (Pb vs. Pv), and thus its rate
gives a more intuitive picture of the basic phys-
of growth will vary with time. If the fissure is
ics involved. Four pressure scales can be defined
close to equidimensional in the fissure plane
(Lister and Kerr, 1991), which represent elasti-
(L ~ B), and buoyancy is absent or small, cracks
city Pe, buoyancy Pb, viscous resistance Pv, and
are usually referred to as “penny-shaped.” Once
fracture toughness, Pf, respectively:
buoyancy has become dominant in the vertical
E H â•… direction and L and B differ significantly (L > B),
Pe ~ , (3.4a)
2(1 − v ) min[ L, B]
2 the n the crack is inherently three dimensional,
and the analysis becomes much harder (Taisne
Pb ~ ∆ρ gL, (3.4b)â•… et al., 2011b). Exercises 3.1–3.3 (see end of chap-
ter) explore the balances that lead to the results
4ηLC â•… presented in Table 3.1, and expose some of con-
Pv ~ , (3.4c)
H2 sequences of assuming two-dimensional geom-
etry in the theory.
Kc â•…
Pf ~ . (3.4d)
min[ L, B]
3.3.2â•… Variable magma and host properties
Here E is Young’s modulus, C is the fissure vel- To describe geological conditions more realis-
ocity scale, and H, B, L are the fissure thickness, tically, certain simplifying assumptions must
breadth, and length scales, respectively, where be discarded. As regards magma properties, the
H and B represent dimensions normal to the most obvious issues are (1) including heat trans-
direction of propagation (with Hâ•›<â•›B), and L is fer in the model will lead to viscosity increases
the dimension in the direction of propagation. and solidification of the magma, particularly
The function min[L,B] means the smaller of the in the tip region, and (2) volatiles initially dis-
two variables is used. The volume of magma in solved in the magma will exsolve to form a gas
the fissure can be specified conveniently in the phase as pressure decreases€ – again the great-
form est effects will be close to the propagating tip.
We have raised several issues that are is injected from a reservoir into the base or
addressed in the discussion of experimental top of a gelatin block. A pre-cut in the gelatin
studies in Section 3.4: (1) Natural fissures have allows clean initiation and orientation of the fis-
finite breadth and therefore a three-dimensional sure. For solidification experiments, the reser-
shape that must be accounted for; only if the voir was kept at a constant temperature above
magma source is linear and very long, may a two- the solidus of the working fluid, while the gel-
dimensional approximation suffice. Of key inter- atin temperature was below this solidus. Visual
est are the factors that might limit their breadth, observations were made by dyeing the fluid
and how they might affect fissure propagation. and viewing normal to the crack plane, using
(2) Laboratory analyses can help to verify the the intensity of color to deduce crack thickness,
scaling laws predicted by theory, such as the l ~ and/or by exploiting the photo-elastic proper-
t1/3 relationship obtained for the case of constant ties of the gelatin to visualize the elastic stress
volume propagation. (3) Experiments can provide field around the fissure. Preparation of the
insights into propagation behavior in the case of gelatin followed the protocol of Menand and
significant variations in magma properties due Tait (2001, 2002)€ – elastic modulus was varied
to degassing and cooling. If a steady regime is via gelatin concentration, and measured in situ
not possible under these conditions, regimes of prior to each experiment. During solidification
unsteady propagation may be defined. experiments the formation of solid caused the
fissure to become opaque, hindering extraction
of quantitative information about fissure thick-
3.4╇ ╇ Experimental investigations ness from the photographs. Nevertheless, fissure
propagation was followed by taking a sequence
3.4.1â•… Experimental methods of photographs at regular intervals, and extract-
Three types of experiments, involving a host ing from the images the total area of the fissure
elastic solid made of gelatin, provide insight as observed normal to the plane of the fissure.
into the dynamics of dike propagation. The first In most natural geologic situations, it is rea-
set of experiments involve continuous injection sonable to suppose that the elastic host medium
of buoyant fluid under isothermal conditions is essentially infinite, whereas this is impossible
into the base of the gelatin, either by maintain- to achieve in the laboratory. Experimental stud-
ing the source reservoir at constant overpres- ies have generally assumed that the experimen-
sure or by imposing a constant injection flux. tal tanks are sufficiently large that wall effects
In a second set of experiments, also isothermal, are negligible, although this does not appear
a constant volume of positively or negatively to have been rigorously tested. Taisne and Tait
buoyant fluid was injected into the gelatin. A (2009, 2011) devised a method of introducing a 2
third set of experiments investigated the effects cm wide vertical layer of water between the gel-
of solidification on propagation, which was atin and the tank walls parallel to the plane of
achieved by injecting hot paraffin at constant the experimental fissure (Fig. 3.9(a), in order to
flux into gelatin at a temperature below the sol- effectively minimize the influence of the finite
idus of the paraffin. The results used for illus- dimensions of the tank on fissure propagation.
tration are drawn from Taisne and Tait (2009,
2011) and Taisne et al. (2011b), but we also dis- 3.4.2â•… Steady propagation regime
cuss work that has produced similar, contrast- The two-dimensional theory of fissure propaga-
ing, or complementary results. tion was first established in the framework of
Experimental techniques and conditions constant imposed flux, but a more natural con-
are described in detail by Taisne and Tait (2009, dition is constant overpressure in the source.
2011), and references therein. In summary, Nevertheless, the calculations of Taisne and
the experiments were carried out in transpar- Jaupart (2009) and experiments of Menand and
ent acrylic tanks 50 cm high and either 30 × Tait (2002) showed that, after an initial transient,
30 cm or 30 × 45 cm in horizontal section. Fluid a steady-state regime is attained for a buoyant
the combined mechanical resistance of gelatin propagation event. The temporal variation of
and frozen crust was sufficient to combat the the crust thickness is the key to understanding
increase in driving pressure the fissure remained the intermittency. Outbreaks commonly spread
static. However, after a time interval that varied around a large part of the dike periphery, and
from experiment to experiment, and to a lesser did not remain confined to or typically even
extent during individual experiments, propaga- initiate at the tip. This mechanism of increas-
tion resumed, accompanied by deflation of the ing the fissure surface area helps to explain the
swollen fissure. The fissure then grew rapidly observed spatial and temporal distribution of
only to be halted again by solidification. seismicity accompanying dike injection events
Measuring the increase in fissure surface (Section 3.2.2; Figs. 3.2, 3.3): much of the seis-
area by analyzing photographs taken at regular micity is generated behind the dike tip during
intervals allowed the rate of fissure propagation intermittent propagation.
to be characterized. Figure 3.13 shows fissure Several other experimental studies also pro-
length and surface area plotted as a function of vide insights into non-steady behaviors observed
time for two experiments with different initial in nature. Menand and Tait (2001) devised a way
conditions. In one case the behavior is quasi- of accumulating gas in the tip region of a propa-
continuous (Fig. 3.13(a)), but in the other case gating liquid-filled crack, and found that once
intermittent propagation is shown by the step- the gas pocket reached a critical size, it was able
wise appearance of the curves (Fig. 3.13(b)). to fracture the gelatin with its own buoyancy.
Figure 3.14(a) shows a photograph of one The gas pocket then propagated faster than
of these experiments just after the dike had the liquid and they proposed this as a possible
reached the surface of the gelatin, and Figure explanation for gas-rich precursor eruptions.
3.14(b) shows a vertical cross section of the upper Muller et al. (2001) carried out experiments in
part of this solidified dike after removal from which a weight was placed on the top of a gel-
the tank. The section features a thick crust that atin block to simulate the presence of a volcanic
formed at the tip during an interval in which edifice. This procedure induced a non-lithostatic
the fissure was stationary, and then the thinner stress field in the gelatin, which in turn caused
crust above that formed during the subsequent dikes to deviate from their previously vertical
trajectory. Acocella and Neri (2009) investigated propagation events, for example, might benefit
propagation within an edifice itself, showing from incorporating some synthesis of theoret-
that the presence of the conical free surface ical models of dike dynamics. This merging of
causes dike trajectories to deviate. Kavanagh the shorter time scale picture into a medium to
et al. (2006) examined propagation through a long-term model of rifts may not be too distant
gelatin block consisting of two superposed lay- a prospect.
ers having different elastic moduli, the upper Two final remarks concern the process of
layer being stiffer. They observed that the dike dike initiation, the mechanisms of which are
would deviate and propagate horizontally at the still not well understood. It is well known
interface between the two layers, suggesting that rock undergoes partial rather than total
that this is one explanation for the formation melting, to produce magma, and the prob-
of sills. lem of how melts are extracted to generate
crystal-poor magmas is an important one. In
the theoretical and experimental work pre-
3.5╇ ╇ Discussion and perspectives sented in this chapter, the boundary condi-
tions concerning dike initiation do not take
Future research on dike propagation is likely to account of the fact that the dike is being fed
take a number of directions, including theoret- from a region that has its own, perhaps slower,
ical, field, and experimental studies, but also dynamic �conditions. Therefore linking of real-
real-time geophysical data collection of active istic magma supply rates or availability with
dike injection events. We have shown how dynamic propagation models is also likely to
laboratory experiments can be used to verify be the subject of interesting future develop-
aspects of theoretical models of dike propagation ments. A few pioneering contributions have
dynamics. This step has been necessary because touched on this matter (Sleep, 1974; Rubin,
of the great difficulty of comparing model pre- 1998), investigating dike behavior in a deform-
dictions directly with field observations. As an able porous medium, but there is a lot of inter-
example of how experiments can also pose a esting work yet to be done. The current view of
challenge to theory, studies including solidifica- magma chambers incorporated in dike propa-
tion demonstrate an intermittent propagation gation models is simplistic in both theory and
regime governed by thermal effects. However, experiment (McLeod and Tait, 1999; Buck et al.,
it remains unclear whether this mechanism is 2006; Rivalta, 2010). It would therefore be of
significant enough to account for the intermit- great interest to devise experiments allowing
tency observed in nature. study of the coupling between a porous source
Several dike propagation events have now and a dike, rather than simply treating an elastic
been recorded by geophysical networks suf- interface.
ficiently dense as to provide quite detailed
information (see Section 3.2.2), considerably
enhancing the potential for making direct com-
parisons of theoretical predictions with the 3.6╇ ╇ Summary
dynamic observations. Geologic observations
of solidified dikes will continue to provide use- • The dynamics of dike emplacement involve
ful insights, but the integration of quantitative complex interactions between the behavior
real-time information from dike-induced seis- of the enclosed fluid and the response of the
mic and deformation measurements is a signifi- host medium.
cant advance. Insights from these new data sets • Regimes of dike behavior can be defined by
have already opened up a collaborative interface analyzing balances between the buoyancy and
with researchers modeling larger-scale tectonic viscous pressures due to the fluid, and elastic
problems. Modeling of divergent boundary tec- and fracture pressures of the host.
tonics at timescales longer than individual dike • Experimental studies have provided insights
Geological Society of London Special Publications, 259, Insights from new data in the recent Manda
43–54, doi:10.1144/GSL.SP.2006.259.01.05 Hararo–Dabbahu rifting episode (Afar, Ethiopia).
Buck, W.R., Einarsson, P. and Brandsdóttir, B. Geochemistry, Geophysics, Geosystems, 12, Q0AB08,
(2006). Tectonic stress and magma chamber size doi:10.1029/2010GC003434.
as controls on dike propagation: Constraints Gudmundsson, A. (2005). The effects of layering and
from the 1975–1984 Krafla rifting episode. local stresses in composite volcanoes on dyke
Journal of Geophysical Research, 111, B12404, emplacement and volcanic hazards. Comptes Rendus
doi:10.1029/2005JB003879. Geosciences, 337, 1216–1222.
Dahm, T. (2000). On the shape and velocity of Gudmundsson, A. and Brenner, S. L. (2001). How
fluid-filled fractures in the Earth. Geophysical Journal hydrofractures become arrested. Terra Nova, 13,
International, 142, 181–192. 456–462.
Delaney, P. T. and Pollard, D. D. (1981). Deformation Gudmundsson, A. and Brenner, S. L. (2005). On
of host rocks and flow of magma during growth the conditions of sheet injections and eruptions
of Minette dikes and breccia-bearing intrusions in stratovolcanoes. Bulletin of Volcanology, 67,
near Ship Rock, New Mexico. United States Geological 768–782.
Survey Professional Paper 1202, 61pp. Halls, H. C. and Fahrig, W. F., eds. (1987). Mafic Dyke
Delaney, P. T. and Pollard, D. D. (1982). Solidification Swarms. Geological Association of Canada Special
of basaltic magma during flow in a dike. American Paper 34.
Journal of Science, 282, 856–885. Hayashi, Y. and Morita, Y. (2003). An image of
Delaney, P. T., Pollard, D. D., Ziony, J. L. and McKnee, magma intrusion process inferred from precise
E. H. (1986). Field relations between dikes and hypocentral migration of the earthquake swarm
joints: Emplacement processes and paleostress east of the Izu Peninsula. Geophysical Journal
analysis. Journal of Geophysical Research, 91, International, 153, 159–174.
4920–4938. Heimpel, M. and Olson, P. (1994). Buoyancy-driven
Ebinger, C. J., Keir, D., Ayele, A. et al. (2008). fracture and magma transport through the
Capturing magma intrusion and faulting processes lithosphere: Models and experiments. In Magmatic
during continental rupture: seismicity of the Systems, ed. M. P. Ryan. San Diego, CA: Academic
Dabbahu (Afar) rift. Geophysical Journal International, Press, pp. 223–240.
174, 1138–1152. Huppert, H. E. and Sparks, R. S. J (1988). The
Ebinger, C. J., Ayele, A., Keir, D. et al. (2010). Length generation of granitic magmas by intrusion of
and timescales of rift faulting and magma intrusion: basalt into continental crust. Journal of Petrology,
The Afar rifting cycle from 2005 to present. Annual 29(3), 599–624.
Review of Earth and Planetary Sciences, 38, 439–466. Jolly, R. and Sanderson, D. (1995). Variation in the
Einarsson, P. and Brandsdóttir, B. (1980). form and distribution of dykes in the Mull swarm,
Seismological evidence for lateral magma Scotland. Journal of Structural Geology, 17, 1543–1557.
intrusion during the July 1978 deflation of the Kavanagh, J. L. and Sparks, R. S. J. (2011). Insights
Krafla volcano in NE-Iceland. Journal of Volcanology of dyke emplacement mechanics from detailed
and Geothermal Research, 47, 160–165. 3D dyke thickness datasets. Journal of the Geological
Ernst, R. E. and Baragar, W. R. A. (1992). Evidence Society, 168, 965–978.
from magnetic fabric for the flow pattern of Kavanagh, J. L., Menand, T. and Sparks, R. S. J. (2006).
magma in the Mackenzie giant radiating dyke An experimental investigation of sill formation
swarm. Nature, 356, 511–513. and propagation in layered elastic media. Earth and
Garagash, D. and Detournay, E. (2000). The tip region Planetary Science Letters, 245, 799–813.
of a fluid-driven crack in an elastic medium. Journal Keir, D., Hamling, I. J., Ayele, A. et al. (2009). Evidence
of Applied Mechanics, 67, 183–192. for focused magmatic accretion at segment centers
Grandin, R., Socquet, A., Binet, R. et al. (2009). from lateral dike injections captured beneath the
September 2005 Manda Hararo-Dabbahu rifting Red Sea rift in Afar. Geology, 37(1), 59–62.
event, Afar (Ethiopia): Constraints provided Lengliné, O., Marsan, D., Got, J., Pinel, V., Ferrazzini,
by geodetic data. Journal of Geophysical Research, V. and Okubo, P. G. (2008). Seismicity and
114(B13), B08404, doi:10.1029/2008JB005843. deformation induced by magma accumulation
Grandin, R., Jacques, E., Nercessian, A. et al. (2011). at three basaltic volcanoes. Journal of Geophysical
Seismicity during lateral dike propagation: Research, 113, B12305, doi:10.1029/2008JB005937.
Lister, J. R. (1990a). Buoyancy-driven fluid fracture: Pinel, V. and Jaupart, C. (2004). Magma storage and
similarity solutions for the horizontal and vertical horizontal dyke injection beneath a volcanic
propagation of fluid-filled cracks. Journal of Fluid edifice. Earth and Planetary Science Letters, 221,
Mechanics, 217, 213–239. 245–262.
Lister, J. R. (1990b). Buoyancy-driven fluid fracture: Pollard, D. D., Delaney, P. T., Duffield, W. A., Endo, E. T.
the effects of material toughness and of and Okamura, A. T. (1983). Surface deformation in
low-viscosity precursors. Journal of Fluid Mechanics, volcanic rift zones. Tectonophysics, 94, 541–584.
210, 263–280. Rivalta, E. (2010). Evidence that coupling to
Lister, J. R. and Kerr, R. C. (1991). Fluid-mechanical magma chambers controls the volume history
models of crack propagation and their application and velocity of laterally propagating intrusions.
to magma transport in dykes. Journal of Geophysical Journal of Geophysical Research, 115, B07203,
Research, 96, 10049–10077. doi:10.1029/2009JB006922.
McLeod, P. and Tait, S. R. (1999). The growth of dykes Roper, S. M. and Lister, J. R. (2007). Buoyancy-driven
from magma chambers. Journal of Volcanology and crack propagation: the limit of large fracture
Geothermal Research, 92, 231–245. toughness. Journal of Fluid Mechanics, 580,
Menand, T. and Tait, S. R. (2001). A 359–380.
phenomenological model for precursor volcanic Rubin, A. M. (1992). Dike-induced faulting and
eruptions. Nature, 411, 678–680. graben subsidence in volcanic rift zones. Journal of
Menand, T. and Tait, S. R. (2002). The propagation of Geophysical Research, 97, 1839–1858.
a buoyant liquid-filled fissure from a source under Rubin, A. M. (1993a). Tensile fracture of rock at
constant pressure: An experimental approach. high confining pressure: Implications for dike
Journal of Geophysical Research, 107(B11), 8471–8481. propagation. Journal of Geophysical Research, 98,
Mériaux, C. and Jaupart, C. (1995). Simple fluid 15,919–15,935.
dynamic models of volcanic rift zones. Earth and Rubin, A. M. (1993b). On the thermal viability of
Planetary Science Letters, 136, 223–240. dikes leaving magma chambers. Geophysical Research
Morita, Y., Nakao, S. and Hayashi, Y. (2006). A Letters, 20(4), 257–260.
quantitative approach to the dike intrusion Rubin, A. M. (1995). Propagation of magma-filled
process inferred from a joint analysis of geodetic cracks. Annual Review of Earth and Planetary Sciences,
and seismological data for the 1998 earthquake 23, 287–336.
swarm off the east coast of Izu Peninsula, central Rubin, A. M. (1998). Dike ascent in partially molten
Japan. Journal of Geophysical Research, 111, B06208, rock. Journal of Geophysical Research, 103(B9),
doi:10.1029/2005JB003860. 20901–20919.
Muller, J. R., Ito, G. and Martel, S. J. (2001). Effects of Rubin, A. M. and Gillard, D. (1998). Dike-induced
volcano loading on dike propagation in an elastic earthquakes: Theoretical considerations. Journal of
half-space. Journal of Geophysical Research, 106(B6), Geophysical Research, 103, 10017–10030.
11101–11113. Rubin, A. M. and Pollard, D. D. (1987). Origins
Muskhelishvili, N. I. (1953). Some Basic Problems of of blade-like dikes in volcanic rift zones. In
the Mathematical Theory of Elasticity. Groningen: P. Volcanism in Hawaii, ed. R. W. Decker, T. L. Wright
Noordhoff Ltd. and P. H. Stauffer, United States Geological Survey
Okada, Y., Yamamoto, E. and Ohkubo, T. (2000). Professional Paper 1350, 1449–1470.
Coswarm and preswarm crustal deformation in Singh, S. C., Crawford, W. C., Carton, H. et al. (2006).
the eastern Izu Peninsula, central Japan. Journal of Discovery of a magma chamber and faults beneath
Geophysical Research, 105, 681–692. a Mid-Atlantic Ridge hydrothermal field. Nature,
Pedersen, R., Sigmundsson, F. and Einarsson, P. 442, 1029–1032.
(2007). Controlling factors on earthquake swarms Sleep, N. H. (1974). Segregation of magma from a
associated with magmatic intrusions: Constraints mostly crystalline mush. Geological Society of America
from Iceland. Journal of Volcanology and Geothermal Bulletin, 85, 1225–1232.
Research, 162, 73–80. Spence, D. A. and Turcotte, D. L. (1985).
Pinel, V. and Jaupart, C. (2000). The effect of edifice Magma-driven propagation of cracks. Journal of
load on magma ascent beneath a volcano. Geophysical Research, 90, 575–580.
Philosophical Transactions of the Royal Society of London, Spence, D. A., Sharp, P. W. and Turcotte, D. L. (1987).
Series A, 358, 1515–1532. Buoyancy-driven crack propagation: a mechanism
for magma migration. Journal of Fluid Mechanics, 174, (a) Derive the buoyancy length scale given by
135–153. Eq. (3.10).
Taisne, B. and Jaupart, C. (2009). Dike propagation (b) Derive the power laws that link length and
through layered rocks. Journal of Geophysical time. (Hint: In the first phase make the
Research, 114, B09203, doi:10.1029/2008JB006228. penny-shape approximation, and in the
Taisne, B. and Jaupart, C. (2011). Magma expansion second phase assume elongation at constant
and fragmentation in a propagating dyke. Earth breadth.)
and Planetary Science Letters, 301, 146–152.
3.2 Consider a penny-shaped crack of fixed volume
Taisne, B. and Tait, S. (2009). Eruption versus
in an elastic solid of the same density. Starting
intrusion? Arrest of propagation of constant
from some initial size, the crack grows because
volume, buoyant, liquid-filled cracks in an elastic,
the stress intensity factor exceeds the fracture
brittle host. Journal of Geophysical Resarch, 114,
toughness of the solid.
B06202, doi:10.1029/2009JB006297.
(a) What is the time dependence of the evo-
Taisne, B. and Tait, S. (2011). The effect
lution of the diameter? (Hint: you should
of solidification on a propagating dike.
derive the appropriate power-law given in
Journal of Geophysical Research, 116, B01206,
Table 3.1.)
doi:10.1029/2009JB007058.
(b) What will be the final static size of the injec-
Taisne, B., Brenguier, F., Shapiro, N. M. and
tion? Take V = 106 m3, E = 109 Pa, Kc = 106 Pa
Ferrazzini, V. (2011a). Imaging the dynamics
m1/2, and ν = 0.5.
of magma propagation using radiated seismic
intensity. Geophysical Research Letters, 38, L04304, 3.3 Figure 3.6 shows the principal balances that
doi:10.1029/2010GL046068. exist within a propagating dike fed by a constant
Taisne, B., Tait, S. and Jaupart, C. (2011b), Conditions flux. From those simple balances you should be
for the arrest of a vertical propagating dyke. able to extract the geometrical characteristics
Bulletin of Volcanology, 73(2), 191–204, doi: 10.1007/ of the problem. The thickness H of the dike, for
s00445–010–0440–1. instance, can be estimated using the equilibrium
Takada, A. (1990). Experimental study on propagation in the tail region, where the viscous pressure
of liquid-filled crack in gelatin: Shape and velocity drop balances the buoyancy, which means that
in hydrostatic stress condition. Journal of Geophysical we can write Pv ~ Pb and replace each pressure by
Research, 95(B6), 8471–8481. their expressions.
Tolstoy, M., Cowen, J. P., Baker, E. T. et al. (2006). (a) What is the expression for H? Compare it
A sea-floor spreading event captured by with Eq. (3.3a).
seismometers. Science, 314, 1920–1922. (b) Next find the typical length L, using the equi-
Weertman, J. (1971). Theory of water-filled crevasses librium that exists within the head region,
in glaciers applied to vertical magma transport Pb ~ Pe.
beneath oceanic ridges. Journal of Geophysical (c) For evaluation of H and L, use reasonable
Research, 76, 1171–1183. order of magnitude values for each param-
Wright, T. J., Ebinger, C., Biggs, J. et al. (2006). Magma- eter: viscosity η = 10 Pa s, density difference
maintained rift segmentation at continental ∆ρ = 100 kg m–3, volumetric flux Qâ•›3D = 106
rupture in the 2005 Afar dyking episode. Nature, m3 s–1, horizontal extent (breadth) B = 10 km,
442, 291–294. Poisson ratio νâ•›=â•›0.5, and Young’s modulus
E = 10 GPa. Discuss the values obtained for
H and L.
Exercises
Online resources available at www.cambridge.
3.1 Consider an experiment in which a fixed �volume org/fagents
of buoyant liquid is released in an elastic
medium. • Answers to exercises
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
5668 − 55.99 pw
Figure 4.1╇╇ Schematic illustration of conduit processes. Cc = pc
T
(
+ 0.4133 pw + 0.002041 p3/2
w
)
.
(a) In effusive eruptions of silicic magma (high viscosity, low
ascent rate) gas may be lost by permeable flow through (4.2)
porous and/or fractured magma. (b) During (sub)plinian
eruptions bubble walls rupture catastrophically at the
Here Cw is total dissolved H2O in wt.%, Cc is dis-
fragmentation surface and the released gas expands rapidly
solved CO2 in ppm and T is temperature in
as the flow changes from a viscous melt with suspended
bubbles to gas with suspended pyroclasts. (c) Extensive loss Kelvin. pw and pc are the partial pressures in
of buoyantly rising bubbles occurs during effusive eruptions MPa of H2O and CO2, respectively. This formu-
of low-volatile content, low-viscosity magma. (d) Coalescence lation is also approximately applicable to other
and accumulation of buoyant bubbles, followed by their melt compositions (Zhang et al., 2007), but more
rupture at the surface, produces strombolian explosions in accurate models are available (Dixon 1997,
slowly ascending low-viscosity magmas. (e) Bubbles remain Newman and Lowenstern 2002, Papale et al.,
coupled to the melt in low-viscosity, hawaiian eruptions, 2006). Equilibrium concentrations of dissolved
which are characterized by relatively high ascent rates,
CO2 and H2O, based on this formulation, are
hydrodynamic fragmentation, and sustained lava fountaining.
shown in Figure 4.2. Notice that almost all CO2
exsolves at >100 MPa for rhyolite and at >25
MPa for basalt, whereas most H2O dissolves at
C w = 0.0012439 p3/2
w <100 MPa for rhyolite and at <25 MPa for basalt.
354.94 pw + 9.623 pw − 1.5223 p3/2 Consequently, processes in the shallow conduit
+ w
(4.1)
T are predominantly affected by H2O exsolution.
(
+ pc −1.084 × 10 −4 p w − 1.362 × 10 −5 p w ) 4.2.2â•… Diffusivity
Volatile diffusivities in silicate melts are best
and characterized for H2O. A recent formulation for
17367 + 1.0964 p m
Dc = −18.239 −
T
+
( 855.2 + 0.2712 pm ) C w
.
T
Diffusive limit
0
During decompression, the volatile concentra-
200 150 100 50 0
tion at the melt–vapor interface, based on the
Ambient pressure, pm (MPa)
assumption of local equilibrium, decreases. This
103
creates a concentration gradient for volatiles to
Fragmentation threshold diffuse to this interface and exsolve. If τdec ≪
(1 MPa/φb) τdif, volatile concentrations will remain close to
102
equilibrium throughout the melt. On the other
hand, if τdif/τdec ≫ 1 (i.e., at high decompression
101
rates) the melt becomes supersaturated, which
pg – pm (MPa)
that brings nearby bubbles into proximity the strain rate in the melt caused by buoyant
(Martula et al., 2000); (3) bubble growth leading rise of the bubble or by magma flow. For small
to stretching and thinning of the liquid film Reb, deformation scales with the Capillary num-
that separates individual bubbles (Borrell and ber, Ca = η0ε̇R/γ, which characterizes the rela-
Leal, 2008); and (4) advection and collision of tive importance of viscous stresses that tend to
bubbles by buoyancy or by magma flow (Manga deform bubbles, and surface tension stresses
and Stone, 1994). that act to keep bubbles spherical. If Ca exceeds
Gravitational and capillary film drainage, some critical value, Cacr, bubble elongation by
as well as bubble collisions are probably most the flow becomes large enough that bubbles
important in low viscosity magmas. A positive will break up. The value of the Cacr depends on
feedback between bubble coalescence, which the steadiness of the flow, melt viscosity, and
increases bubble size, and bubble mobility due flow type. For the viscosities of silicate melts,
to buoyancy is expected (Section 4.3.5). Whereas Cacr should range from ~1 to >103. In both types
capillary drainage is dominant at bubble radii of flows, bubbles will become highly elongated
<3â•›cm, gravitational drainage becomes dom- before breakup occurs. Capillary numbers will
inant at radii >3 cm (Proussevitch et al., 1993). generally be large enough in conduits for large
A characteristic velocity for capillary drainage deformation and breakup to occur. However, to
can be obtained by balancing the capillary pres- achieve large deformation also requires large
sure gradient, γ/R2, with viscous resistance to strains which may be limited to the sides of
flow, η0vf/R2, where vf is the velocity in the film. conduits. It is thought that tube pumice is a
Because the bubble–melt surface is assumed to manifestation of these conditions (Marti et al.,
be a free-slip surface, the characteristic length 1999).
scale is the bubble radius R rather than film In basaltic magmas, inertial forces can no
thickness. The resultant scaling vf ~ γ/η0 high- longer be neglected once bubbles become larger
lights the importance of melt viscosity. Whether than a few centimeters. In this limit, velocity
and at what rate bubble collisions result in differences across bubbles can lead to breakup.
coalescence depends on the Weber number, The relative importance of inertial forces and
We = 2ρv2R/γ, where v is the velocity at which surface tension forces is characterized by We,
two bubbles approach one another. based on the velocity difference across the melt
In high viscosity magmas bubbles remain that surrounds the bubble. For large Reb, breakup
essentially “frozen” in the melt and coalescence occurs if We exceeds a critical value, Wecr, which
is principally a consequence of bubble growth. implies that there might be a maximum stable
In this case, coalescence appears to create a per- bubble size (Hinze, 1955). Although the actual
meable network of bubbles due to relatively value of Wecr depends on the origin of the vel-
persistent holes in the ruptured melt films (Klug ocity differences, it is generally between 1 and 5
and Cashman, 1996). Because the interfacial ten- for a wide range of flow conditions.
sion between melt and crystals tends to be lower
than between gas and crystals, coalescence in 4.3.5â•… Bubble mobility
the presence of crystals is expected to occur at Bubbles are buoyant and their rise speed depends
greater film thickness than in the absence of on size, volume fraction, and viscosity. If buoy-
crystals (Proussevitch et al., 1993b). ant bubble rise is much slower than magma
ascent, bubbles are dispersed throughout the
4.3.4â•… Breakup continuous melt phase and move passively with
The breakup of bubbles into several smaller the flow. Such flows are called “dispersed” and
bubbles originates with their deformation in the asymptotic limit of infinitesimally small
by some combination of viscous stresses and dispersed bubbles the flow is termed “homoge-
inertial forces. In silicic magmas, the bubble neous.” If the velocity of buoyant bubble rise is
Reynolds number, Reb = ε̇R2 ρ/η0, will always be similar to or greater than magma ascent veloci-
small and inertia can be neglected. Here ε̇ is ties, bubbles are decoupled from the liquid phase
and Ghiorso, 2000). For highly viscous magmas equation (Carman, 1956) where permeability, k,
the flow below the fragmentation level will, is given as
under most conditions, be laminar (Re < 103) and
( )
k (φb) = χ φb − φbp .
β
f � 16/Re. Above the fragmentation level the flow (4.15)
will be turbulent and f � f0. For basaltic magma,
frictional pressure loss over a wide range of Here χ is an empirical constant and ϕbp is the
ascent velocities is less than magma-static pres- volume fraction of bubbles that corresponds to
sure loss and the flow dynamics will be more percolation threshold (Saar and Manga, 1999),
sensitive to changes in the density of the ascend- with typical values of 2 � β � 4. An uncontro-
ing magma than for eruptions where frictional versial permeability model for vesicular magma
pressure losses are larger. Consequently, under remains elusive (Takeuchi et al., 2005; Wright
separated flow conditions there is potential et al., 2009). Furthermore, it appears that the
for feedbacks between discharge rate, gas-to- nature of magma deformation plays a critical
melt flux ratios, and flow regime (Seyfried and role in creating permeability (Okumura et al.,
Freundt, 2000; Guet and Ooms, 2006). 2008; Okumura et al., 2010).
Permeable outgassing can be modeled as
4.3.7â•… Permeable outgassing flow through a porous medium by combining
The degassing history during ascent of an indi- the continuity equation for the exsolved volatile
vidual magma parcel may be complex and phase with Darcy’s equation or, if the velocity
in many cases gas may separate from rising of the gas phase becomes too large for inertial
parcels of magma. We refer to this as outgas- effects to be neglected, Forcheimer’s equation
sing, which may be a consequence of buoyant (Rust and Cashman, 2004).
bubble rise, magma fragmentation, or perme-
able gas flow. Chapters 6 and 8 discuss outgas-
sing associated with separated flow and high
gas-to-melt flux ratios during strombolian and
4.4╇ ╇Crystal nucleation and
perhaps hawaiian eruptions. In some silicic growth
eruptions high melt-to-gas flux ratios are also
suggestive of decoupled gas flow (Edmonds Crystals nucleate due to undercooling, which
et al., 2003). However, high viscosities and neg- is predominantly a consequence of H2O exsolu-
ligible bubble mobility require that outgassing tion. Nucleation rates determine the crystal
is associated with magma permeability, pre- number density and size distribution, which
sumably a consequence of coalescing bubbles provide a record of the magma’s ascent history.
that form a permeable network (Eichelberger During crystallization the volatile content of the
et al., 1986; Klug and Cashman,1996) and per- residual melt phase increases because volatiles
haps cracks and fractures produced by brittle preferentially partition into the melt phase. This
magma deformation (Gonnermann and Manga, in turn affects bubble nucleation and growth, as
2003; Gonnermann and Manga, 2005a; Tuffen well as melt viscosity.
et al., 2003). It has been suggested that if con- Undercooling, the thermodynamic driving
duit walls are permeable, which is controver- force for crystallization, is the chemical poten-
sial (Boudon et al., 1998), then volatiles may tial of H2O in the melt and can be expressed as
also escape laterally from ascending magma �Te, the difference between actual temperature,
into the conduit walls and permeable outgas- T, and the liquidus temperature, Tl. Because
sing has the potential to modulate eruptive the latter depends on composition and vola-
behavior (Jaupart and Allegre, 1991; Woods and tile content, bubble nucleation and growth will
Koyaguchi, 1994; Eichelberger, 1995). increase Tl and can result in sufficient undercool-
Vesicularity-permeability measurements for ing for crystallization (Hammer, 2004). The rate
volcanic rocks, and by inference magmas, are of homogeneous nucleation of critical nuclei, I
usually analyzed using the Kozeny–Carman (m–3 s–1), is given by classical nucleation theory as
∆G * + ∆G D
I = I 0 exp − (4.16) 4.5╇ ╇ Magma rheology
k BT .
Silicate melts form a disordered network of
Here I0 is a reference nucleation rate, kB is the interconnected SiO4 tetrahedra where the
Boltzmann constant, and �GD is the activation self-diffusive motion of atoms, called structural
energy for diffusion (James, 1985). �G* is the relaxation, results in a continuous unstructured
free energy required to form a spherical nucleus rearrangement of the molecular structure. In
of critical size and depends on the free energy, the presence of an applied stress, this molecular
σ, associated with the crystal–liquid interface. rearrangement results in a directional motion
Therefore σ, nucleation rate increases with of SiO4 tetrahedra relative to one another and
decreasing σ (Mueller et al., 2000). The rate of macroscopically manifests itself as viscous flow
heterogeneous nucleation also follows Eq. (4.16), (Moynihan, 1995).
but with a modified σ to account for the lower The intrinsic viscosity of silicate melts varies
interfacial energies (Spohn et al., 1988). Because over orders of magnitude, even within a single
σ is relatively poorly constrained, estimates of volcanic eruption. It depends on the degree of
crystal nucleation rates are typically associated polymerization, a function of chemical com-
with large uncertainties. Moreover, the validity position and volatile content. Within realistic
of classical nucleation theory and its assump- ranges of compositional variability, eruptive
tions that (1) the interface between nucleus and temperature, and volatile content, the viscosity
melt is sharp; (2) the critical nucleus has the of basaltic melts may vary between about 10 and
thermodynamic properties of the bulk solid; 103 Pa s, and between about 104 and 1012 Pa s or
and (3) σ can be treated as a macroscopic prop- more for silicic melts (Fig. 4.5). Changes in vis-
erty equal to the value for a planar interface, cosity during individual eruptions are especially
remain controversial. pronounced at low pressures (depths <1â•›km),
Once a crystal nucleates its growth rate, Y, is where most of the H2O exsolves (Figs. 4.2 and
given by (Spohn et al., 1988) 4.6). Magma rheology is discussed further in the
context of lava flow emplacement in Chapter 5.
∆H ∆Te ∆G D
Y = Y0 1 − exp (4.17)
k BTl T
exp − k T ,
B 4.5.1â•…The effect of dissolved volatiles
and temperature
where Y0 is a reference growth rate and �H is Dissolved water dissociates into molecular
the change in enthalpy between melt and crys- water and hydroxyl ions, thereby depolymer-
talline phases. If empirical parameters such izing the melt. In rhyolitic melts the effect of
as Tl and σ are known, the above equations water content can be tremendous and exceeds
can be used to model magma crystallization. the effect of temperature (Fig. 4.5), but it is less
Alternatively, parameterizations for I and Y can pronounced for mafic magmas (Giordano and
be used to explore the dependence of eruption Dingwell, 2003). In contrast, viscosity is less
behavior on syneruptive crystallization (Melnik affected by CO2 (Bourgue and Richet, 2001).
and Sparks, 1999). The prediction of crystal size Recent viscosity formulations applicable over a
distributions is possible through the use of the range of compositions, temperature, and water
Avrami equation (Marsh, 1998) contents are discussed in Hui and Zhang (2007),
Zhang et al. (2007), and Giordano et al. (2008).
( )
φx = 1 − exp − k v Y 3 t 4 I . (4.18)
4.5.2â•… The effect of deformation rate
Here a given crystal volume fraction, ϕx, is If the applied rate of deformation exceeds a cer-
assumed to correspond to thermodynamic equi- tain threshold, the induced molecular motions
librium after crystallization over time, t, and of the melt are no longer compensated by ran-
with kv as a volumetric factor. dom reordering of the melt structure. This is
G∞
ε st ∼ 10 −3 . (4.19)
η0
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Crystal content, φx
Figure 4.8╇╇The effect of bubbles at small and large Ca
Figure 4.7╇╇The effect of crystals on the relative mixture on the velocity profile in a cylindrical conduit for a given
viscosity (ηr = η/η0), based on Eq. (4.20), with parameters pressure gradient. Dashed curve is the bubble-free reference
ϕbcr =€–0,066499 tanh (0.913424 log(ε̇) + 3:850623) case Ca = 1 (or equivalently ϕb = 0). Both solid curves are for
+ 0:591806; ϕb = 0.6 and ϕbcr = 0.75 in Eq. (4.21).
δ╇ =€–6.301095 tanh (0.818496 log(ε̇) + 2.86) + 7.462605;
α╇ =€–0.000378 tanh (1.148101 log(ε̇) + 3.92) + 0.999572;
β╇ = 3.987815 tanh (0.8908 log (ε̇) + 3.24) + 5.099645. 1999; Stein and Spera, 2002). An expression for
this dependence on Ca is given by Pal (2003) as
not involve any interaction with non-magmatic Namiki and Manga (2005) suggested a poten-
water. The type and efficiency of the fragmenta- tial energy criterion, based on the observation
tion process determines pyroclast size and how that during rapid decompression ϕb and �pf
much magmatic gas is released per unit mass of determine the expansion velocity of a bubbly
magma. This, in turn, has implications for the liquid. Potential energy depends on both ϕb and
volcanic jet, column, and plume that are pro- �pf (Mastin, 1995), with fragmentation taking
duced when the gas-pyroclast mixture exits the place above some threshold.
volcanic vent (Chapters 8 and 9). Fragmentation Hydrodynamic or inertial fragmentation should
is thought to occur when specific conditions be predominantly associated with mafic magma,
are reached within the conduit. The proposed where rapid decompression results in rapid bub-
criteria for fragmentation include: (1) a critical ble growth, inertial stretching, and hydrodynamic
bubble volume fraction; (2) a stress criterion; (3) breakup (Shimozuru, 1994; Zimanowski et al.,
a strain-rate criterion; (4) a potential energy cri- 1997; Villermaux, 2007). A criterion for inertial
terion; and (5) an inertial criterion. fragmentation is based on the Reynolds number,
The critical volume fraction criterion is Rev, and Weber number, Wev, of the expanding
thought to arise from some form of instabil- magma (Namiki and Manga, 2008)
ity within the thin bubble walls, once ϕb ≈ 0.75
is reached (Verhoogen, 1951; Sparks, 1978).
ρ (1 − φb) v 2 L
Rev = > O (1) (4.24)
However, magma fragments have vesicularities 3η0
that range from 0 (obsidian) to >98% (reticulite)
implying that a fragmentation criterion gov- and
erned by a critical volume fraction cannot be
ρ (1 − φ b) v 2 L
generally applicable (though very high vesicu- Wev = 1. (4.25)
larities may reflect post-fragmentation growth γ
of bubbles).
The stress criterion is based on the view Here L is a characteristic length scale and v
that fragmentation in high-viscosity magmas is the expansion velocity of the vesicular
takes place when volatile overpressure, �pf, magma, which can be obtained from the bubble
exceeds the tensile strength of the melt and growth rate.
ruptures bubble walls (Alidibirov, 1994; Zhang,
1999). Spieler et al. (2004) provide a formula-
tion with good fit to a broad range of experimen- 4.7╇ ╇ Modeling of magma ascent
tal data
The past decade has seen much progress in
∆p f = 106 Pa / φb, (4.23) modeling magma ascent (Sahagian, 2005). In
contrast to the modeling of effusive eruptions,
that has been modified by Mueller et al. (2008) to explosive eruptions require a criterion to esti-
account for permeable gas flow. mate the depth of magma fragmentation, and
The strain-rate criterion is based on the merging of a model for the bubbly flow region
observation that silicate melts can fragment if below the fragmentation level with one for
deformation rates exceed the structural relax- the gas-pyroclast region above. Flow in the
ation rate of the melt at ε̇ ~ 102 ε̇st (Section 4.5.2). gas-pyroclast flow region is compressible and
It is thought to be the consequence of a rapid is typically modeled as one-dimensional with a
decompression that forces rapid bubble growth parameterization for wall friction and granular
(Papale, 1999). Ittai et al. (2010) have argued that stresses during highly turbulent flow (Wilson
both the stress criterion and the strain-rate cri- and Head, 1981; Dobran, 1992; Koyaguchi,
terion are equivalent, because of their mutual 2005).
dependency on shear modulus and magma Conduit models may be steady or
rheology. time-dependent. The steady approximation is
4.7.1â•…Steady homogeneous flow in one Expanding Eq. (4.26) and substituting for du/dz
dimension in Eq. (4.27) gives
Below the fragmentation level homogeneous dp f ρ u 2 dA dρ
flow models advect bubbles passively with the − = ρ g + ρu2 − − u2 . (4.28)
melt. That is Ut ≪ Um and u = Um = Ub, where dz a A dz dz
Ub is the bubble velocity. Together with the
assumption of constant mass flux throughout Equation (4.28) indicates that the change
the conduit, Q, the equation of mass conserva- in magma pressure with respect to height
tion becomes depends, from left to right, on magma-static
pressure loss, frictional pressure loss, change in adjust the flow back to 1 atm at the vent exit.
conduit diameter, and change in magma dens- If the exit pressure exceeds 1 atm, the flow is
ity. Assuming isentropic conditions (denoted by “choked,” that is M = 1 and the exit velocity
subscript S), that is an infinitesimal reversible is equal to c, the speed of sound in the gas–
pressure change and negligible time for heat pyroclast mixture.
transfer, then In many cases the assumption of isothermal
conditions is reasonable; if not, an equation for
dρ ∂ρ dp 1 dp conservation of energy has to be added (Mastin
= = 2 . (4.29)
dz ∂p S dz c dz and Ghiorso, 2000; Dobran, 2001)
Here c is the speed of sound in the magma and dH + udu + gdz = 0. (4.32)
using the Mach number, M = u2/c2, Eq. (4.28)
becomes Here H is the specific enthalpy of the melt–gas
mixture from which the magma temperature
can be calculated.
dp f ρu 2 dA
−
dz
( )
1 − M 2 = ρ g + ρu 2 −
a A dz
. (4.30)
du g ∂u
dp ∂u
ρg u g φb = − φb − Fmg − Fwg − ρφb g (4.35) τ rz = − η z + r . (4.41)
dz dz ∂r ∂z
and
du m dp Here λ = 2 η − K , where η is the viscosity of the
ρm u m (1 − φb) = −(1 −φb ) − Fmg − Fwm 3
dz dz (4.36)
− ρm (1 − φb ) g. magma. K is called the bulk viscosity, which
accounts for the compressibility of the two-phase
magma mixture (Massol et al., 2001) and is often
Here Fwg and Fwm are the wall drag forces for the
assumed negligible. Various simplifications of
phase that is assumed to be in contact with the
Eqs. (4.37) and (4.38) can be employed. If the flow
conduit wall. For example, Dobran (1992) used
is laminar (Re < 103), which is usually the case
Fwm = (dp/dz)η, Fwg = 0 below the fragmentation
below the fragmentation level, the second and
level and Fwm = 0, Fwg = (dp/dz)η above.
third terms on the left-hand side can be neglected.
Furthermore, if the flow is steady the time deriv-
4.7.3â•… Two-dimensional flow atives are zero. Assuming isothermal conditions,
If magma flow is radially non-uniform, a
the momentum equations are complemented by
two-dimensional modeling approach can provide
an equation for the conservation of mass
additional insight. Cases where two-dimensional
modeling has been important involve ∂ρ 1 ∂ ∂
non-Newtonian magma rheology and shear heat-
+ ( ρ rur ) + ( ρ uz ) = 0. (4.42)
∂t r ∂r ∂z
ing, as well as permeable outgassing and heat Examples of recent two-dimensional conduit
loss through the conduit walls. models are given by Collier and Neuberg (2006);
The direct approach is to solve the Navier– Costa et al. (2007a,b); Hale (2007); Hale and
Stokes equations in two cartesian or axisymmet- Muehlhaus (2007).
ric cylindrical coordinates. For axisymmetric If vertical flow is a reasonable approximation
homogeneous flow, conservation of momentum (u = uz, ur = 0), Eqs. (4.37) and (4.38) become
is given by (Bird et al., 1960)
∂ ∂p
∂u ∂u ∂u ∂r
( r τ rz ) = − r ,
∂z
(4.43)
ρ r + ρu r r + ρ u z r
∂t ∂r ∂z
(4.37)
∂p 1 ∂ τ rz
=− − ( r τ rr ) − and upon integration
∂r r ∂r z
∂u z r ∂p
and τ rz = − η =− . (4.44)
∂r 2 ∂z
∂u z ∂u ∂u
ρ + ρur z + ρuz z
∂t ∂r ∂z The relative ease with which Eq. (4.44) can be
(4.38)
∂p 1 ∂ ∂τ solved makes it easier to model more complex
= −ρg − −
∂z r ∂r
( r τ rz ) − zz ,
∂z magma rheologies (Costa and Macedonio, 2005;
Mastin, 2005; Vedeneeva et al., 2005), and to include
with stress tensors subgrid scale calculations of bubble growth,
where the model for bubble growth is discretized
∂u r 1 ∂ ∂u at a smaller scale than the conduit model itself
τ rr = −2η − λ ( ru r ) + z , (4.39)
∂r r ∂r ∂z (Fig. 4.9; Gonnermann and Manga, 2003, 2007).
To investigate the role of shear heating
∂u z 1 ∂ ∂u
τ zz = −2η − λ
r ∂r
( rur ) + z , (4.40) or magma cooling near the conduit walls, an
∂z ∂z energy equation has to be included. Neglecting
thermal diffusion in the vertical direction, the
and energy equation is (Bird et al., 1960)
∆pf
3
Depth (km)
(pg-pm)
8
0 100 200 0 0.5 1 10–2 100 102 104 106 108 10–3 10–1 101
–1
pm (MPa) φb ∆p (MPa) η0 (Pa s) dT/dt (K s )
�
behavior (Melnik and Sparks,1999; Macedonio discharge rates on the order of 106 kg s–1 (sub-
et al., 2005). plinian) and 106–108 kg s–1 (plinian), sustained
over several hours (Chapter 8). Syneruptive vola-
tile exsolution and crystallization can lead to
4.8╇ ╇What conduit models viscosity-limited bubble growth, large frictional
pressure losses, and large decompression rates
have taught us (Papale and Dobran, 1993, 1994), resulting in gas
overpressure, supersaturation, and bubble nucle-
Figure 4.10 shows the relationship between ation (Massol and Koyaguchi, 2005), as well as
magma composition (in essence a proxy for fragmentation (Papale, 1999).
viscosity), eruption rate (or equivalently ascent Whether fragmentation takes place depends
rate), and eruption style. The boundaries shown on the conditions of magma ascent. However,
are not “hard” in that uncertainties in thresholds fragmentation also feeds back on eruption
for fragmentation, unsteady flow, variation in dynamics. For example, the presence of pum-
volatile content, conduit geometry, two-dimen- ice fragments after fragmentation, as opposed
sional flow, and crystallization are not addressed. to complete fragmentation to ash, means that
Overall these boundaries are meant to provide not all compressed gas contained in bubbles
insight into the conditions under which certain may be released, resulting in ascent velocities
processes are expected to dominate and to illus- and pressures that are lower than if all gas were
trate their relationships with eruption style. released. Incomplete outgassing of coarse frag-
ments also increases the density of the erupting
4.8.1â•… Subplinian and plinian eruptions gas–pyroclast mixture, with consequences for
Subplinian and plinian eruptions are character- the dynamics of the eruption column (Chapter 8,
ized by brittle magma fragmentation at mass Papale, 2001). Furthermore, the granular stress
induced by >~1 cm-size pyroclasts decreases and Martini (2009); James et al. (2009); Pioli
exit velocities and enhances the lateral inhomo- et al. (2009).
geneity in clast-size distribution throughout the
conduit (Dufek and Bergantz, 2005). 4.8.3â•… Hawaiian eruptions
The strong dependence of viscosity on tem- Hawaiian eruptions are characterized by gas-rich
perature in silicic magmas is of considerable fountains of basaltic magma sustained for hours
importance for shear heating near the conduit to days at mass discharge rates of the order of 104
walls (Rosi et al., 2004; Polacci et al., 2005). The to >106 kg s–1 (Houghton and Gonnermann, 2008).
resulting viscosity reduction will change the Fountains reach heights that exceed hundreds of
flow profile toward a more plug-like flow and meters producing ash and pyroclasts, presumably
reduce frictional pressure loss. Consequently, through hydrodynamic fragmentation (Namiki
the discharge rate increases for a given pressure and Manga, 2008). Microtextural studies of pyro-
drop between chamber and surface, and hence clasts indicate less open-system gas loss than for
affects the depth of fragmentation and allows strombolian clasts (Polacci et al., 2006).
for multiple stable discharge rates (Costa and Conduit models by Wilson and Head (1981)
Macedonio, 2002; Mastin, 2005; Vedeneeva et al., predict that magma ascent in hawaiian eruptions
2005; Costa et al., 2007). is homogeneous, with dispersed bubbles, so that
Eruption intensity may correlate with pre- rapid decompression and shallow H2O exsolution
eruptive H2O content and magma composition lead to expansion-driven acceleration and the for-
(Papale et al., 1998; Polacci et al., 2004; Starostin mation of a sustained magma fountain above the
et al., 2005). However, changes in CO2 content vent (Parfitt et al., 1995; Parfitt, 2004). The tran-
can have the opposite effect of H2O (Papale and sition from hawaiian to strombolian behavior is
Polacci, 1999; Ongaro et al., 2006). The transition thought to occur at magma rise speeds of less
from explosive to effusive eruptions in silicic than about 1 m s–1, with a correlation between
magmas is thought to occur as magma ascent required rise speed and initial magmatic volatile
rates decrease, which allows more time for per- content (Parfitt and Wilson, 1995). Alternatively,
meable outgassing and bubbles grow at near based on analog laboratory experiments it has
equilibrium conditions (Jaupart and Allegre, been suggested that bubbles accumulate to form
1991; Woods and Koyaguchi, 1994; Gonnermann a magmatic foam, whose collapse produces
and Manga, 2007). an annular gas-melt flow within the conduit
(Chapter 6; Jaupart and Vergniolle, 1988).
4.8.2â•… Strombolian eruptions
Typical strombolian eruptions (Chapter 6) 4.8.4â•… Effusive eruptions
consist of prolonged periods of impulsive Large volumes of basaltic magma often erupt
explosions that typically eject 0.01–10â•›m3 of effusively over prolonged periods of time, at
pyroclastic material and volcanic gases at about rates up to 104 kg s–1 (Wolfe et al., 1987). Silicic
100 m s–1. Mass balance considerations imply magmas of high viscosity may erupt effusively to
that much of the gas is derived from a larger vol- form lava domes and coulees with time-averaged
ume of magma than is erupted, a consequence eruption rates ranging from 10–1 to 104 kg s–1
of slow magma ascent rates (<â•›0.01–0.1 m s–1). (Pyle, 2000). The resulting low ascent rates
This permits significant bubble coalescence to (� 0.01 m s–1) are thought to facilitate permeable
form gas slugs and/or accumulations of bubbles, outgassing (Eichelberger et al., 1986; Jaupart
which decouple from the melt and rise to the and Allegre, 1991; Woods and Koyaguchi, 1994;
surface where they burst. Aside from thermal Massol and Jaupart, 1999). Sparks (1997) attrib-
modeling (Giberti et al., 1992), some numerical uted the occurrence of time-dependent behav-
conduit models that have dealt directly with ior during dome-forming eruptions to feedbacks
details of strombolian eruptions are Wilson and associated with volatile-dependent magma vis-
Head (1981); Parfitt and Wilson (1995); James cosity, magma degassing, outgassing, and crys-
et al. (2008); O’Brian and Bean (2008); D’Auria tallization. Conduit models suggest that large
changes in discharge rate at relatively small subaerial flow, and incorporation of bubble
changes in boundary conditions are the conse- nucleation.
quence of such feedbacks (Melnik and Sparks, (2) The dynamics of magma flow within the
1999, 2005). They also demonstrate how non- volcanic conduit during strombolian and
linear feedbacks produce different steady dis- hawaiian eruptions remains controversial.
charge rates for the same boundary conditions The main challenge here is the modeling of
but different initial conditions, perhaps explain- the separated two-phase flow, where the top-
ing how transitions between effusive and explo- ology of the flow can change drastically and
sive behavior occur (Melnik et al., 2005; Clarke rapidly in both space and time.
et al., 2007). (3) Magma flow in the volcanic conduit is
For long-lived dome-forming eruptions, affected by processes that occur over a
magma recharge is often treated as constant wide range of length scales, from the fluid
over part of the eruption. However, transient dynamics of magma ascent over kilometers
eruptive behavior can ensue as a direct con- to crystal/bubble nucleation and growth,
sequence of compressibility of the vesicular as well as fragmentation over millimeter
magma and/or surrounding wall rock (Meriaux to micrometer scales. Accounting for these
and Jaupart, 1995; Denlinger, 1997; Huppert different scales requires the development
and Woods, 2002; Costa et al., 2007a). Feedbacks of subgrid-type models capable of coupling
may be such that pressure build-up drives the explicit modeling of the various small-scale
system into a state where discharge rate exceeds processes with fluid dynamical modeling of
recharge rate. With time, pressure decreases magma flow.
until the cycle repeats or is dynamically damp- (4) Eruption triggering is perhaps of foremost
ened toward a stable state (Barmin et al., 2002). importance in terms of hazard mitigation.
High shear stress at the conduit walls, in con- Integration of geodetic and remote-sensing
junction with shear-thinning rheology, may observations with models of magma flow,
result in shear localization and brittle deform- supply, and storage is thus highly desirable.
ation and this may in turn enhance permeabil- A major challenge is the incorporation of
ity and outgassing. Conduit models have also geological (tectonic and structural) com-
explored how brittle deformation may be asso- plexities, which may require a departure
ciated with stick-slip behavior at the conduit for idealized one- or two-dimensional model
walls and shallow seismicity, especially for mag- geometries, as well as coupling between
mas with high crystal content (Denlinger and tectonically produced stresses and stresses
Hoblitt, 1999; Goto, 1999; Collier and Neuberg, within the magmatic system.
2006; Iverson et al., 2006; Hale, 2007).
Di diffusivity of dissolved volatile species i Qin magma chamber mass recharge rate
(m2 s–1) (kg s–1)
Dwb diffusivity of H2O dissolved in basalt Qout magma chamber mass discharge rate
melt (m2 s–1) (kg s–1
Dwr diffusivity of H2O dissolved in rhyolite Qm mass flow rate of melt phase (kg s–1)
melt (m2 s–1) r radial distance from center of bubble or
f friction factor for flow in a cylindrical from center of conduit (m)
pipe R bubble radius (m)
Fwg wall drag force of the gas phase (N) S radial distance from bubble center to
Fwm wall drag force of the melt phase (N) midpoint of between adjacent
g acceleration due to gravity (m s–2) bubbles (m)
G∞ shear modulus of melt (Pa) t time (s)
�GD activation energy for diffusion (J) T temperature (K)
�G* free energy for nucleus formation (J) Tl liquidus temperature (K)
h hindering function for Richardson–Zaki �Te undercooling of melt (K)
equation u magma velocity (m s–1)
H specific enthalpy (J kg–1) ug gas velocity (m s–1)
�H change in enthalpy between melt and um melt velocity (m s–1)
crystal (J) ur radial component of magma velocity
I homogeneous crystal nucleation rate (m s–1)
(m–3 s–1) uz vertical component of magma velocity
I0 reference homogeneous crystal (m s–1)
nucleation rate (m–3 s–1) Ub bubble ascent velocity (m s–1)
J bubble nucleation rate (m–3 s–1) Um ascent velocity of the melt phase (m s–1)
k permeability (m2) Ut terminal rise velocity of a single bubble
k v volumetric factor for Avrami equation in infinite liquid (m s–1)
kB Boltzmann constant (J kg–1) v approach velocity of two bubbles or
K bulk viscosity (Pa s) expansion velocity of vesicular magma
L characteristic length scale for inertial (m s–1)
fragmentation (m) vf velocity in the melt film between
n empirical coefficient bubbles (m s–1)
N d number of bubbles per volume of vr radial velocity of the melt surrounding
melt (m–3) a growing bubble (m s–1)
p magma pressure (Pa) vR radial velocity of the melt-vapor
pc partial pressure of CO2 (MPa) interface of growing bubble (m s–1)
pm pressure of the melt (Pa) V volume of magma chamber (m3)
ṗm melt decompression rate (Pa s–1) Y crystal growth rate (m s–1)
p g pressure of the vapor inside Y0 reference crystal growth rate (m s–1)
bubbles (Pa) z vertical coordinate (m)
pw partial pressure of H2O (MPa) α empirical constant
�pf critical gas overpressure for brittle β empirical constant for Kozeny–Carman
fragmentation (Pa) equation
�ps supersaturation pressure of volatile in βe effective compressibility (m2 N–1)
melt (Pa) γ vapor–melt surface tension (N m–1)
Q mass flow rate of magma (kg s–1) δ empirical constant
Q g mass flow rate of gas phase (kg s–1) ε strain rate (s–1)
Qgd mass flow rate of dissolve volatile phase ε̇f critical strain rate for structural failure
(kg s–1) of the melt (s–1)
ε̇st strain rate at which shear thinning the manuscript. Both authors were supported by
begins (s–1) NSF while preparing this review.
η magma viscosity (Pa s)
η0 Newtonian melt (liquid) viscosity (Pa s)
ηg viscosity of vapor inside a bubble (Pa s) References
ηr Relative viscosity
θ contact angle Alidibirov, M. (1994). A model for viscous magma
κ Thermal conductivity (W K–1 m–1) fragmentation during volcanic blasts. Bulletin of
Volcanology, 56, 459–465.
ρ magma density (kg m–3)
Allard, P., Burton, M. and Mure, F. (2005).
ρ g vapor density (kg m–3)
Spectroscopic evidence for a lava fountain driven
ρ m melt density (kg m–3)
by previously accumulated magmatic gas. Nature,
σ free energy of crystal–liquid interface (J)
433, 407–410.
τdec characteristic decompression timescale (s) Arbaret, L., Bystricky, M. and Champallier, R. (2007).
τdif characteristic diffusion time of volatile Microstructures and rheology of hydrous synthetic
in melt (s) magmatic suspensions deformed in torsion at high
τif components of the stress tensor (Pa) pressure. Journal of Geophysical Research, 112.
τr viscous relaxation time (s) Bagdassarov, N. S. and Dingwell, D. B. (1992). A
τvis characteristic viscous timescale (s) rheological investigation of vesicular rhyolite.
ϕb volume fraction of bubbles Journal of Volcanology and Geothermal Research, 50,
ϕbp percolation threshold 307–322.
ϕx volume fraction of crystals Bagdassarov, N., Dorfman, A. and Dingwell, D. B.
ϕxcr critical volume fraction of crystals (2000). Effect of alkalis, phosphorus, and water on
ϕbcr critical volume fraction of bubbles the surface tension of haplogranite melt. American
χ empirical constant for Kozeny–Carman Mineralogist, 85, 33–40.
Barmin, A., Melnik, O. and Sparks, R. S. J. (2002).
equation
Periodic behavior in lava dome eruptions. Earth and
ψ geometrical factor
Planetary Science Letters, 199, 173–184.
Ca Capillary number, η0ε̇R/γ
Batchelor, G. K. (1967). An Introduction to Fluid
Cacr Capillary number at which bubbles Dynamics. Cambridge: Cambridge University
break up, η0ε̇R/γ Press.
M Mach number, u2/c2 Behrens, H. and Zhang, Y. X. (2001). Ar diffusion in
Re Reynolds number for flow in a hydrous silicic melts: implications for volatile
cylindrical pipe, auz/η diffusion mechanisms and fractionation. Earth and
Reb Reynolds number for a growing bubble, Planetary Science Letters, 192, 363–376.
ε̇R2ρ/η0 Bird, R. B., Steward, W. E. and Lightfoot, E. N. (1960).
Rev Reynolds number of expanding magma, Transport Phenomena. New York: John Wiley.
ρ(1 – ϕb) v2L/(3η0) Blank, J. G. and Brooker, R. A. (1994). Experimental
We Weber number, 2ρv2R/γ studies of carbon-dioxide in silicate melts;
Wecr Weber number at which bubbles break solubility, speciation, and stable carbon-isotope
up, 2ρv2R/γ behavior. In M. R. Carroll and J. R. Holloway, (Eds.)
Volatiles in Magmas, American Mineralogical Society,
Wev Weber number of expanding magma,
30 Reviews in Mineralogy, 157–186.
ρ(1 – ϕb)v2L/γ
Bluth, G. J. S., Schnetzler, C. C., Krueger, A. J. and
Walter, L. S. (1993). The contribution of explosive
volcanism to global atmospheric sulfur-dioxide
Acknowledgments concentrations. Nature, 366, 327–329.
Borrell, M. and Leal, L. G. (2008). Viscous coalescence
The authors thank B. F. Houghton for stimulating of expanding low-viscosity drops; the dueling
discussions, as well as J. E. Hammer, L. Karlstrom, drops experiment. Journal of Colloid and Interface
and M. Rudolph for constructive comments on Science, 319, 263–269.
Boudon, G., Villemant, B., Komorowski, J. C., Costa, A., Melnik, O. and Vedeneeva, E. (2007).
Ildefonse, P. and Semet, M. P. (1998). The Thermal effects during magma ascent in conduits.
hydrothermal system at Soufriere Hills volcano, Journal of Geophysical Research, 112.
Montserrat (West Indies): Characterization and D’Auria, L. and Martini, M. (2009). Slug flow:
role in the on-going eruption. Geophysical Research Modeling in a conduit and associated elastic
Letters, 25, 3693–3696. radiation. In R. A. Meyers, (Ed.) Monitoring and
Bourgue, E. and Richet, P. (2001). The effects of mitigation of volcano hazards, New York: Springer,
dissolved CO2 on the density and viscosity of 8153–8168.
silicate melts: a preliminary study. Earth and Denlinger, R. P. (1997). A dynamic balance between
Planetary Science Letters, 193, 57–68. magma supply and eruption rate at Kilauea
Brennen, C. E. (2005). Fundamentals of Multiphase volcano, Hawaii. Journal of Geophysical Research, 102,
Flow. New York: Cambridge University Press, 2nd 18091–18100.
edition. Denlinger, R. P. and Hoblitt, R. P. (1999). Cyclic
Bruce, P. M. and Huppert, H. E. (1989). Thermal eruptive behavior of silicic volcanoes. Geology, 27,
control of basaltic fissure eruptions. Nature, 342, 459–462.
665–667. Dixon, J. E. (1997). Degassing of alkalic basalts.
Caricchi, L., Burlini, L., Ulmer, P., et al. (2007). American Mineralogist, 82, 368–378.
Non-Newtonian rheology of crystal-bearing Dobran, F. (1992). Nonequilibrium flow in volcanic
magmas and implications for magma ascent conduits and application to the eruptions of Mt.
dynamics. Earth and Planetary Science Letters, 264, St. Helens on May 18, 1980, and Vesuvius in A.D.
402–419. 79. Journal of Volcanology and Geothermal Research, 49,
Carman, P. C. (1956). Flow of Gases Through Porous 285–311.
Media. San Diego, CA: Academic. Dobran, F. (2001). Volcanic Processes Mechanisms in
Carmichael, I. S. E. (2002). The andesite aqueduct: Material Transport. New York, N Y: Kluwer Academic
perspectives on the evolution of intermediate / Plenum Publishers.
magmatism in west-central (105–99 degrees W) Dufek, J. and Bergantz, G. W. (2005). Transient
Mexico. Contributions to Mineralogy Petrology, 143, two-dimensional dynamics in the upper conduit
641–663. of a rhyolitic eruption: A comparison of closure
Carroll, M. R. and Webster, J. D. (1994). Solubilities of models for the granular stress. Journal of Volcanology
sulfur, noble-gases, nitrogen, chlorine, and fluorine and Geothermal Research, 143, 113–132.
in magmas. Volatiles in Magmas, 30, 231–279. Edmonds, M., Oppenheimer, C., Pyle, D. M., Herd, R.
Clarke, A. B., Stephens, S., Teasdale, R., Sparks, R. A. and Thompson, G. (2003). SO2 emissions from
S. J. and Diller, K. (2007). Petrologic constraints Soufriere Hills Volcano and their relationship to
on the decompression history of magma prior conduit permeability, hydrothermal interaction
to Vulcanian explosions at the Souffiere Hills and degassing regime. Journal of Volcanology and
volcano, Montserrat. Journal of Volcanology and Geothermal Research, 124, 23–43.
Geothermal Research, 161, 261–274. Eichelberger, J. C. (1995). Silicic volcanism. Annual
Collier, L. and Neuberg, J. (2006). Incorporating Reviews in Earth and Planetary Science, 23, 41–63.
seismic observations into 2D conduit flow Eichelberger, J. C., Carrigan, C. R., Westrich, H.
modeling. Journal of Volcanology and Geothermal R. and Price, R. H. (1986). Non-explosive silicic
Research, 152, 331–346. volcanism. Nature, 323, 598–602.
Costa, A. and Macedonio, G. (2005). Viscous heating Eposti Ongaro, T., Papale, P., Neri, A. and Del Seppia,
effects in fluids with temperature-dependent D. (2006). Influence of carbon dioxide on the
viscosity: Triggering of secondary flows. Journal of large-scale dynamics of magmatic eruptions at
Fluid Mechanics, 540, 21–38. Phlegrean Fields (Italy). Geophysical Research Letters,
Costa, A., Melnik, O. and Sparks, R. S. J. (2007). 33.
Controls of conduit geometry and wallrock Gardner, J. E. and Denis, M. H. (2004). Heterogeneous
elasticity on lava dome eruptions. Earth and bubble nucleation on Fe-Ti oxide crystals
Planetary Science Letters, 260, 137–151. in high-silica rhyolitic melts. Geochimica et
Costa, A., Melnik, O., Sparks, R. S. J. and Voight, B. Cosmochimica Acta, 68, 3587–3597.
(2007). Control of magma flow in dykes on cyclic Gardner, J. E. (2007). Heterogeneous bubble
lava dome extrusion. Geophysical Research Letters, 34. nucleation in highly viscous silicate melts during
instantaneous decompression from high pressure. Hammer, J. E. (2004). Crystal nucleation in hydrous
Chemical Geology, 236, 1–12. rhyolite: Experimental data applied to classical
Ghiaasiaan, S. M. (2008). Two-Phase Flow, Boiling, and theory. American Mineralogist, 89, 1673–1679.
Condensation in Conventional and Miniature Systems. Hauri, E. (2002). SIMS analysis of volatiles in silicate
New York: Cambridge University Press. glasses, 2: isotopes and abundances in Hawaiian
Giberti, G. and Wilson, L. (1990). The influence of melt inclusions. Chemical Geology, 183, 115–141.
geometry on the ascent of magma in open fissures. Hinze, J. O. (1955). Fundamentals of the
Bulletin of Volcanology, 52, 515–521. hydrodynamic mechanisms of splitting in
Giberti, G., Jaupart, C. and Sartoris, G. (1992). dispersion processes. American Institute of Chemical
Steady-state operation of Stromboli Volcano, Engineers Journal, 1, 289–295.
Italy: constraints on the feeding system. Bulletin of Hirth, J. P., Pound, G. M. and St. Pierre, G. R. (1970).
Volcanology, 54, 535–541. Bubble nucleation. Met allurgical and Materials
Giordano, D. and Dingwell, D. B. (2003). Viscosity Transactions B, 1, 939–945.
of hydrous Etna basalt: implications for Houghton, B. F. and Gonnermann, H. M. (2008).
Plinian-style basaltic eruptions. Bulletin of Basaltic explosive volcanism: constraints from
Volcanology, 65, 8–14. deposits and models. Chemie der Erde€– Geochemistry,
Giordano, D., Russell, J. K. and Dingwell, D. B. (2008). 68, 117–140.
Viscosity of magmatic liquids: A model. Earth and Hui, H. J. and Zhang, Y. X. (2007). Toward a general
Planetary Science Letters, 271, 123–134. viscosity equation for natural anhydrous and
Gonnermann, H. M. and Manga, M. (2003). Explosive hydrous silicate melts. Geochimica et Cosmochimica
volcanism may not be an inevitable consequence Acta, 71, 403–416.
of magma fragmentation. Nature, 426, 432–435. Huppert, H. E. and Woods, A. W. (2002). The role of
Gonnermann, H. M. and Manga, M. (2005a). Flow volatiles in magma chamber dynamics. Nature,
banding in obsidian: A record of evolving textural 420, 493–495.
heterogeneity during magma deformation. Earth Hurwitz, S. and Navon, O. (1994). Bubble nucleation
and Planetary Science Letters, 236, 135–147. in rhyolitic melts: Experiments at high-pressure,
Gonnermann, H. M. and Manga, M. (2005b). temperature, and water content. Earth and Planetary
Nonequilibrium magma degassing: Results from Science Letters, 122, 267–280.
modeling of the ca.1340 AD eruption of Mono Ittai, K., Lyakhovsky, V. and Navon, O. (2010). Bubble
Craters, California. Earth and Planetary Science Letters, growth in visco-elastic magma: Implications to
238, 1–16. magma fragmentation and bubble nucleation.
Gonnermann, H. M. and Manga, M. (2007). The fluid Bulletin of Volcanology, 73, 39–54.
mechanics inside a volcano. Annual Reviews in Fluid Iverson, R. M., Dzurisin, D., Gardner, C. A. et
Mechanics, 39, 321–356. al. (2006). Dynamics of seismogenic volcanic
Gonnermann, H. M. and Mukhopadhyay, S. (2007). extrusion at Mount St Helens in 2004–05. Nature,
Non-equilibrium degassing and a primordial 444, 439–443.
source for helium in ocean-island volcanism. James, M. R., Lane, S. J. and Corder, S. B. (2008).
Nature, 449, 1037–1040. Modelling the rapid near-surface expansion of gas
Goto, A. (1999). A new model for volcanic earthquake slugs in low viscosity magmas. In S. J. Lane and
at Unzen Volcano: Melt rupture model. Geophysical J. S. Gilbert, (Eds.) Fluid Motion in Volcanic Conduits:
Research Letters, 26, 2541–2544. A Source of Seismic and Acoustic Signals, Geological
Guet, S. and Ooms, G. (2006). Fluid mechanical Society of London Special Publication, 307,
aspects of the gas-lift technique. Annual Reviews in 147–167.
Fluid Mechanics, 38, 225–249. James, M. R., Lane, S. J., Wilson, L. and Corder,
Hale, A. J. (2007). Magma flow instabilities in a S. B. (2009). Degassing at low magma-viscosity
volcanic conduit: Implications for long-period volcanoes: Quantifying the transition between
seismicity. Physics of the Earth and Planetary Interiors, passive bubble-burst and Strombolian eruption.
163, 163–178. Journal of Volcanology and Geothermal Research, 180,
Hale, A. J. and Muehlhaus, H. B. (2007). Modelling 81–88.
shear bands in a volcanic conduit: Implications James, P. F. (1985). Kinetics of crystal nucleation in
for over-pressures and extrusion-rates. Earth and silicate glasses. Journal of Non-Crystalline Solids, 73,
Planetary Science Letters, 263, 74–87. 517–540.
Jaupart, C. and Allegre, C. J. (1991). Gas content, flow. Journal of Volcanology and Geothermal Research,
eruption rate and instabilities of eruption regime 143, 205–217.
in silicic volcanoes. Earth and Planetary Science Macedonio, G., Dobran, F. and Neri, A. (1994).
Letters, 102, 413–429. Erosion processes in volcanic conduits and
Jaupart, C. and Vergniolle, S. (1988). Laboratory application to the AD 79 eruption of
models of Hawaiian and Strombolian eruptions. Vesuvius. Earth and Planetary Science Letters, 121,
Nature, 331, 58–60. 137–152.
Johnson, M. C., Anderson, A. T. and Rutherford, Macedonio, G., Neri, A., Marti, J. and Folch, A. (2005).
M. J. (1994). Pre-eruptive volatile contents of Temporal evolution of flow conditions in sustained
magmas. In Volatiles In Magmas, Washington D. C.: magmatic explosive eruptions. Journal of Volcanology
Mineralogical Society of America, 30 Reviews In and Geothermal Research, 143, 153–172.
Mineralogy, 281–330. Manga, M., Castro, K. V., Cashman, K. V. and
Kennedy, B., Spieler, O., Scheu, B. et al.(2005). Loewenberg, M. (1998). Rheology of bubble-bearing
Conduit implosion during Vulcanian eruptions. magmas: Theoretical results. Journal of Volcanology
Geology, 33, 581–584. and Geothermal Research, 87, 15–28.
Kerrick, D. M. and Jacobs, G. K. (1981). A modified Manga, M. and Stone, H. A. (1994). Interactions
Redlich-Kwong Equation for H2O, CO2, and H2O-CO2 between bubbles in magmas and lavas: Effects
mixtures at elevated pressures and temperatures. of bubble deformation. Journal of Volcanology and
American Journal of Science, 281, 735–767. Geothermal Research, 63, 267–279.
Klug, C. and Cashman, K. V. (1996). Permeability Mangan, M. and Sisson, T. (2000). Delayed,
development in vesiculating magmas: implications disequilibrium degassing in rhyolite magma:
for fragmentation. Bulletin of Volcanology, 58, Decompression experiments and implications
87–100. for explosive volcanism. Earth and Planetary Science
Koyaguchi, T. (2005). An analytical study for Letters, 183, 441–455.
1-dimensional steady flow in volcanic conduits. Mangan, M. and Sisson, T. (2005). Evolution
Journal of Volcanology and Geothermal Research, 143, of melt-vapor surface tension in silicic
29–52. volcanic systems: Experiments with hydrous
Lange, R. A. (1994). The effect of H2O, CO2 and F on melts. Journal of Geophysical Research, 110,
the density and viscosity of silicate melts. Revies in doi:10.1029/2004JB003215.
Mineralogy, 30, 331–369. Marsh, B. D. (1998). On the interpretation of crystal
Lavallee, Y., Hess, K. U., Cordonnier, B. and Dingwell, size distributions in magmatic systems. J. Petrol.,
D. B. (2007). Non-Newtonian rheological law 39, 553–599.
for highly crystalline dome lavas. Geology, 35, Marti, J., Soriano, C. and Dingwell, D. B. (1999). Tube
843–846. pumices as strain markers of the ductile-brittle
Lejeune, A. M., Bottinga, Y., Trull, T. W. and Richet, P. transition during magma fragmentation. Nature,
(1999). Rheology of bubble-bearing magmas. Earth 402, 650–653.
and Planetary Science Letters, 166, 71–84. Martula, D. S., Hasegawa, T., Lloyd, D. R. and
Lejeune, A. M. and Richet, P. (1995). Rheology of Bonnecaze, R. T. (2000). Coalescence-induced
crystal-bearing silicate melts€– An experimental coalescence of inviscid droplets in a viscous
study at high viscosities. Journal of Geophysical fluid. Journal of Colloid and Interface Science, 232,
Research, 100, 4215–4229. 241–253.
Lensky, N. G., Navon, O. and Lyakhovsky, V. (2004). Massol, H. and Jaupart, C. (1999). The generation of
Bubble growth during decompression of magma: gas overpressure in volcanic eruptions. Earth and
experimental and theoretical investigation. Journal Planetary Science Letters, 166, 57–70.
of Volcanology and Geothermal Research, 129, 7–22. Massol, H., Jaupart, C. and Pepper, D. W. (2001).
Liu, Y., Zhang, Y. X. and Behrens, H. (2005). Solubility Ascent and decompression of viscous vesicular
of H2O in rhyolitic melts at low pressures and a magma in a volcanic conduit. Journal of Geophysical
new empirical model for mixed H2O-CO2 solubility Research, 106, 16223–16240.
in rhyolitic melts. Journal of Volcanology and Massol, H. and Koyaguchi, T. (2005). The effect
Geothermal Research, 143, 219–235. of magma flow on nucleation of gas bubbles
Llewellin, E. W. and Manga, A. (2005). Bubble in a volcanic conduit. Journal of Volcanology and
suspension rheology and implications for conduit Geothermal Research, 143, 69–88.
Mastin, L. G. (1995). Thermodynamics of gas and model written in Visual Basic for Excel. Comp.
steam-blast eruptions. Bulletin of Volcanology, 57, Geosci., 28, 597–604.
85–98. Nowak, M., Schreen, D. and Spickenbom, K. (2004).
Mastin, L. G. (2005). The controlling effect of viscous Argon and CO2 on the race track in silicate melts:
dissipation on magma flow in silicic conduits. Journal A tool for the development of a CO2, speciation
of Volcanology and Geothermal Research, 143, 17–28. and diffusion model. Geochimica et Cosmochimica
Mastin, L. G. and Ghiorso, M. S. (2000). A numerical Acta, 68, 5127–5138.
program for steady-state flow of magma-gas O’Brian, G. S. and Bean, C. J. (2008). Seismicity
mixtures through vertical eruptive conduits. on volcanoes generated by gas slug ascent.
United States Geological Survey, 00–209 Open-File Geophysical Research Letters, 35, L16308, doi:
Report, 53. 10.1029/2008gl035001.
McMillan, P. F. (1994). Water solubility and speciation Okumura, S., Nakamura, M., Nakano, T.,
models. In M. R. Carroll and J. R. Holloway, (Eds.) Uesugi, K. and Tsuchiyama, K. (2010). Shear
Volatiles in Magmas, Mineralogical Society of deformation experiments on vesicular rhyolite:
America, 30 Reviews in Mineralogy, 131–156. Implications for brittle fracturing, degassing,
Melnik, O., Barmin, A. A. and Sparks, R. S. J. (2005). and compaction of magmas in volcanic
Dynamics of magma flow inside volcanic conduits conduits. Journal of Geophysical Research, 115,
with bubble overpressure buildup and gas loss doi:10.1029/2009JB006904.
through permeable magma. Journal of Volcanology Okumura, S., Nakamura, M., Tsuchiyama, K., Nakano,
and Geothermal Research, 143, 53–68. T. and Uesugi, K. (2008). Evolution of bubble
Melnik, O. and Sparks, R. S. J. (1999). Nonlinear microstructure in sheared rhyolite: Formation of a
dynamics of lava dome extrusion. Nature, 402, 37–41. channel-like bubble network. Journal of Geophysical
Melnik, O. and Sparks, R. S. J. (2005). Controls on Research, 113, doi:10.1029/2007JB005362.
conduit magma flow dynamics during lava Pal, R. (2003). Rheological behavior of bubble-bearing
dome building eruptions. Journal of Geophysical magmas. Earth and Planetary Science Letters, 207,
Research, 110. 165–179.
Meriaux, C. and Jaupart, C. (1995). Simple fluid Papale, P. (1999). Strain-induced magma fragmentation
dynamic models of volcanic rift zones. Earth and in explosive eruptions. Nature, 397, 425–428.
Planetary Science Letters, 136, 223–240. Papale, P. (2001). Dynamics of magma flow in
Mitchell, K. L. (2005). Coupled conduit flow and volcanic conduits with variable fragmentation
shape in explosive volcanic eruptions. Journal of efficiency and nonequilibrium pumice degassing.
Volcanology and Geothermal Research, 143, 187–203. Journal of Geophysical Research, 106, 11043–11065.
Moynihan, C. T. (1995). Structural relaxation and Papale, P. (2005). Determination of total H2O and CO2
the glass transition. Revies in Mineralogy and budgets in evolving magmas from melt inclusion
Geochemistry, 32, 1–19. data. Journal of Geophysical Research, 110.
Mueller, S., Scheu, B., Spieler, O. and Dingwell, Papale, P. and Dobran, F. (1993). Modeling of the
D. B. (2008). Permeability control on magma ascent of magma during the Plinian eruption
fragmentation. Geology, 36, 399–402. of Vesuvius in AD 79. Journal of Volcanology and
Muller, R., Zanotto, E. D. and Fokin, V. M. (2000). Geothermal Research, 58, 101–132.
Surface crystallization of silicate glasses: Papale, P. and Dobran, F. (1994). Magma flow along
nucleation sites and kinetics. Journal of the volcanic conduit during the Plinian and
Non-Crystalline Solids, 274, 208–231. pyroclastic flow phases of the May 18, 1980,
Namiki, A. and Manga, M. (2005). Response of a bubble Mount-St. Helens eruption. Journal of Geophysical
bearing viscoelastic fluid to rapid decompression: Research, 99, 4355–4373.
Implications for explosive volcanic eruptions. Earth Papale, P., Moretti, R. and Barbato, D. (2006). The
and Planetary Science Letters, 236, 269–284. compositional dependence of the saturation
Namiki, A. and Manga, M. (2008). Transition between surface of H2O+CO2 fluids in silicate melts. Chemical
fragmentation and permeable outgassing of Geology, 229, 78–95.
low viscosity magmas. Journal of Volcanology and Papale, P., Neri, A. and Macedonio, G. (1998). The role of
Geothermal Research, 169, 48–60. magma composition and water content in explosive
Newman, S. and Lowenstern, J. B. (2002). eruptions€– 1. Conduit ascent dynamics. Journal of
VOLATILECALC: A siliate melt-H2O-CO2 solution Volcanology and Geothermal Research, 87, 75–93.
Papale, P. and Polacci, M. (1999). Role of carbon Journal of Volcanology and Geothermal Research, 59,
dioxide in the dynamics of magma ascent in 161–178.
explosive eruptions. Bulletin of Volcanology, 60, Pyle, D. M. (2000). Sizes of volcanic eruptions. In
583–594. B. F. Houghton, H. Rymer, J. Stix, S. McNutt and
Parfitt, E. A. (2004). A discussion of the mechanisms H. Sigurdsson, (Eds.) Encyclopedia of Volcanoes, San
of explosive basaltic eruptions. Journal of Volcanology Diego: Academic Press, 263–269.
and Geothermal Research, 134, 77–107. Pyle, D. M. and Pyle, D. L. (1995). Bubble migration
Parfitt, E. A. and Wilson, L. (1995). Explosive and the initiation of volcanic eruptions. Journal of
volcanic eruptions€– IX. The transition between Volcanology and Geothermal Research, 67, 227–232.
Hawaiian-style lava fountaining and Strombolian Richardson, J. F. and Zaki, W. N. (1954). The
explosive activity. Geophysical Journal International, sedimentation of a suspension of uniform
121, 226–232. spheres under conditions of viscous flow. Chemical
Parfitt, E. A., Wilson, L. and Neal, C. A. (1995). Factors Engineering Science, 3, 65–73.
influencing the height of Hawaiian lava fountains: Rosi, M., Landi, P., Polacci, M., Di Muro, A. and
Implications for the use of fountain height as Zandomeneghi, D. (2004). Role of conduit shear
an indicator of magma gas content. Bulletin of on ascent of the crystal-rich magma feeding the
Volcanology, 57, 440–450. 800-year-BP Plinian eruption of Quilotoa Volcano
Pioli, L., Azzopardi, B. J. and Cashman, K. V. (Ecuador). Bulletin of Volcanology, 66, 307–321.
(2009). Controls on the explosivity of scoria Rosner, D. E. and Epstein, M. (1972). Effects of
cone eruptions: Magma segregation at conduit interface kinetics, capillarity and solute diffusion
junctions. Journal of Volcanology and Geothermal on bubble growth rates in highly supersaturated
Research, 186, 407–415. liquids. Chemical Engineering Science, 27, 69–88.
Polacci, M., Corsaro, R. A. and Andronico, D. Rust, A. C. and Cashman, K. V. (2004). Permeability
(2006). Coupled textural and compositional of vesicular silicic magma: Inertial and hysteresis
characterization of basaltic scoria: Insights into effects. Earth and Planetary Science Letters, 228, 93–107.
the transition from Strombolian to fire fountain Rust, A. C., Manga, M. and Cashman, K. V. (2003).
activity at Mount Etna, Italy. Geology, 34, 201–204. Determining flow type, shear rate and shear stress
Polacci, M., Papale, P., Del Seppia, D., Giordano, D. in magmas from bubble shapes and orientations.
and Romano, C. (2004). Dynamics of magma ascent Journal of Volcanology and Geothermal Research, 122,
and fragmentation in trachytic versus rhyolitic 111–132.
eruptions. Journal of Volcanology and Geothermal Saar, M. O. and Manga, M. (1999).
Research, 131, 93–108. Permeability-porosity relationship in vesicular
Polacci, M., Rosi, M., Landi, P., Di Muro, A. and basalts. Geophysical Research Letters, 26, 111–114.
Papale, P. (2005). Novel interpretation for shift Sahagian, D. (2005). Volcanic eruption mechanisms:
between eruptive styles in some volcanoes. Insights from intercomparison of models of
Eos, Transactions, American Geophysical Union, 86, conduit processes. Journal of Volcanology and
333–336. Geothermal Research, 143, 1–15.
Press, W. H., Teukolsky, S. A., Vetterling, W. T. and Sahagian, D. L. and Proussevitch, A. A. (1996).
Flannery, B. P. (1992). Numerical Recipes in C: The Thermal effects of magma degassing. Journal of
Art of Scientific Computing. Cambridge: Cambridge Volcanology and Geothermal Research, 74, 19–38.
University Press. Scriven, L. E. (1959). On the dynamics of phase
Proussevitch, A., Sahagian, D. and Anderson, A. growth. Chemical Engineering Science, 10, 1–13.
(1993a). Dynamics of diffusive bubble growth in Seyfried, R. and Freundt, A. (2000). Experiments on
magmas: Isothermal case. Journal of Geophysical conduit flow and eruption behavior of basaltic
Research, 98, 22,283–22,307. volcanic eruptions. Journal of Geophysical Research,
Proussevitch, A. A. and Sahagian, D. L. (1998). 105, 23727–23740.
Dynamics and energetics of bubble growth in Shimozuru, D. (1994). Physical parameters governing
magmas: Analytical formulation and numerical the formation of Pele’s hair and tears. Bulletin of
modeling. Journal of Geophysical Research, 103, Volcanology, 56, 217–219.
18223–18251. Slezin, Y. B. (2003). The mechanism of volcanic
Proussevitch, A. A., Sahagian, D. L. and Kutolin, V. A. eruptions (a steady state approach). Journal of
(1993b). Stability Of Foams In Silicate Melts. Volcanology and Geothermal Research, 122, 7–50.
Sparks, R. S. J. (1978). The dynamics of bubble Toramaru, A. (1989). Vesiculation process and
formation and growth in magmas: A review bubble-size distributions In ascending magmas
and analysis. Journal of Volcanology and Geothermal with constant velocities. Journal of Geophysical
Research, 3, 1–37. Research, 94, 17523–17542.
Sparks, R. S. J. (1997). Causes and consequences of Tuffen, H., Dingwell, D. B. and Pinkerton, H. (2003).
pressurisation in lava dome eruptions. Earth and Repeated fracture and healing of silicic magma
Planetary Science Letters, 150, 177–189. generate flow banding and earthquakes? Geology,
Sparks, R. S. J., Barclay, J., Jaupart, C., Mader, H. M. 31, 1089–1092.
and Phillips, J. C. (1994). Physical aspects of Vedeneeva, E. A., Melnik, O. E., Barmin, A. A. and
magmatic degassing I. Experimental and Sparks, R. S. J. (2005). Viscous dissipation in
theoretical constraints on vesiculation. In Volatiles explosive volcanic flows. Geophysical Research
In Magmas, Washington D. C.: Mineralogical Society Letters, 32.
of America, 30 Reviews In Mineralogy, 413–445. Verhoogen, J. (1951). Mechanics of ash formation.
Spieler, O., Kennedy, B., Kueppers, U. et al. (2004). The American Journal of Science, 249, 729–739.
fragmentation threshold of pyroclastic rocks. Earth Villermaux, E. (2007). Fragmentation. Annual Reviews
and Planetary Science Letters, 226, 139–148. in Fluid Mechanics, 39, 419–446.
Spohn, T., Hort, M. and Fischer, H. (1988). Wallace, P. J. (2005). Volatiles in subduction
Numerical simulation of the crystallization of zone magmas: concentrations and fluxes
multicomponent melts in thin dikes or sills. 1. The based on melt inclusion and volcanic gas data.
liquidus phase. Journal of Geophysical Research, 93, Journal of Volcanology and Geothermal Research, 140,
4880–4894. 217–240.
Starostin, A. B., Barmin, A. A. and Melnik, O. E. Wallace, P. J., Carn, S. A. R., William, I., Bluth, G. J. S.
(2005). A transient model for explosive and and Gerlach, T. M. (2003). Integrating petrologic
phreatomagmatic eruptions. Journal of Volcanology and remote sensing perspectives on magmatic
and Geothermal Research, 143, 133–151. volatiles and volcanic degassing. Eos, Transactions,
Stein, D. J. and Spera, F. J. (2002). Shear viscosity American Geophysical Union, 84, 446–447.
of rhyolite-vapor emulsions at magmatic Wallis, G. B. (1969). One-Dimensional Two-phase Flow.
temperatures by concentric cylinder rheometry. McGraw–Hill.
Journal of Volcanology and Geothermal Research, 113, Watson, E. B. (1994). Diffusion in volatile-bearing
243–258. magmas. Revies in Mineralogy, 30, 371–411.
Stevenson, R. J., Dingwell, D. B., Webb, S. L. and Webb, S. (1997). Silicate melts: Relaxation, rheology,
Sharp, T. G. (1996). Viscosity of microlite-bearing and the glass transition. Reviews of Geophysics, 35,
rhyolitic obsidians: An experimental study. Bulletin 191–218.
of Volcanology, 58, 298–309. Wilson, L. and Head, J. W. (1981). Ascent and
Tait, S., Jaupart, C. and Vergniolle, S. (1989). Pressure, eruption of basaltic magma on the Earth and
gas content and eruption periodicity of a shallow, Moon. Journal of Geophysical Research, 86, 2971–3001.
crystallizing magma chamber. Earth and Planetary Wilson, L., Sparks, R. S. J. and Walker, G. P. L. (1980).
Science Letters, 92, 107–123. Explosive volcanic eruptions€– IX. The control
Takeuchi, S., Nakashima, S., Tomiya, A. and of magma properties and conduit geometry on
Shinohara, H. (2005). Experimental constraints on eruption column behavior. Geophysical Journal of the
the low gas permeability of vesicular magma during Royal Astronomical Society, 63, 117–148.
decompression. Geophysical Research Letters, 32. Wolfe, E. W., Garcia, M. O., Jackson, D. B. et al.
Taylor, G. I. (1932). The viscosity of a fluid containing (1987). The Pu’u O’o eruption of Kilauea Volcano,
small drops of another fluid. Proceedings of the episodes 1–20, January 3, 1983, to June 8, 1984. In
Royal Society of London A: Mathematical, Physical & Decker, R. W. and Wright, T. L. and Stauffer, P. H.,
Engineering Sciences, 138, 41–48. (Ed.) Volcanism in Hawaii, United States Geological
Thies, M. (2002). Herstellung und rheologische Survey, 1350 United States Geological Survey
Eigenschaften von porösen Kalk-Natron-Silicatschmelzen. Professional Paper, 471–508.
Ph. d., RU. Woods, A. W., Bokhove, O., de Boer, A. and Hill,
Thomas, N., Jaupart, C. and Vergniolle, S. (1994). On B. E. (2006). Compressible magma flow in a
the vesicularity of pumice. Journal of Geophysical two-dimensional elastic-walled dike. Earth and
Research, 99, 15633–15644. Planetary Science Letters, 246, 241–250.
Woods, A. W. and Cardoso, S. S. S. (1997). Triggering Zimanowski, B., Buttner, R., Lorenz, V. and Hafele,
basaltic volcanic eruptions by bubble-melt H. G. (1997). Fragmentation of basaltic melt in the
separation. Nature, 385, 518–520. course of explosive volcanism. Journal of Geophysical
Woods, A. W. and Koyaguchi, T. (1994). Transitions Research, 102, 803–814.
between explosive and effusive eruption of silicic Zimanowski, B., Wohletz, K., Dellino, P. and Buttner,
magmas. Nature, 370, 641–644. R. (2003). The volcanic ash problem. Journal of
Wright, H. M. N., Cashman, K. V., Gottesfeld, E. Volcanology and Geothermal Research, 122, 1–5.
H. and Roberts, J. J. (2009). Pore structure of
volcanic clasts: Measurements of permeability and
electrical conductivity. Earth and Planetary Science Exercises
Letters, 280.
Wylie, J. J. and Lister, J. R. (1995). The effects of Consider two magmas, a basalt (η = 1 Pa s, ρ = 2800 kg
temperature-dependent viscosity on flow in a m–3) and rhyolite (η = 106 Pa s, ρ = 2600 kg m–3), rising
cooled channel with application to basaltic fissure in a conduit. Assume a conduit radius of 10 m and 1
eruptions. Journal of Fluid Mechanics, 305, 239–261. m for the rhyolite and basalt, respectively. Assume a
Wylie, J. J., Voight, B. and Whitehead, J. A. (1999). pressure gradient driving magma ascent of 500 Pa m–1.
Instability of magma flow from volatile-dependent For both cases:
viscosity. Science, 285, 1883–1885.
Zenit, R., Koch, D. L. and Sangani, A. S. (2001). 4.1 Calculate the velocity as a function of radial pos-
Measurements of the average properties of a ition in the conduit.
suspension of bubbles rising in a vertical channel. 4.2 Evaluate the strain rates at the conduit walls.
Journal of Fluid Mechanics, 429, 307–342. Are the strain rates large enough that the melt
Zhang, Y. X. (1999). A criterion for the fragmentation might be shear-thinning or undergo structural
of bubbly magma based on brittle failure theory. failure?
Nature, 402, 648–650. 4.3 Calculate the capillary number for a 1 cm radius
Zhang, Y. X. and Behrens, H. (2000). H2O diffusion in bubble. Will the bubbles become deformed by
rhyolitic melts and glasses. Chemical Geology, 169, the ascending magma?
243–262. 4.4 Calculate the rise speed of a 1 cm bubble relative
Zhang, Y. X., Xu, Z. J., Zhu, M. F. and Wang, H. Y. to the surrounding melt. Given this speed, what
(2007). Silicate melt properties and volcanic flow regime is likely to characterize the magma
eruptions. Reviews of Geophysics, 45. (see Fig. 4.4).
Lava flows
Andrew J. L. Harris
Overview and link lava flow heat loss, cooling rates, crys-
tallinity, rheology, velocity, and thickness.
There are many types of lava flow emplace-
ment models. One type of numeric model aims
to apply and link equations that describe lava
flow cooling rates, rheology, and dynamics to
5.1╇ ╇Lava flows and lava flow
simulate lava flow spreading. Generally applied models
to basaltic flow emplacement, it is this type of
model that we consider here. Within this model Following the classic lava channel morphological
type, we can define two simulation sub-types: model of Lipman and Banks (1987), an active
volume- and cooling-limited. By applying and lava flow system can be split into two principal
linking laws governing the ability of a Bingham components: a feeder system, which is either a
fluid to spread under the influence of gravity, channel or tube, and a zone of dispersed lava
both volume- and cooling-limited simulations flow supplied by the feeder system. Within this
spread control volumes across a digital top- system, a range of lava types, defined by their
ography. Volume-limited approaches spread characteristic surface morphologies (including
and thin a finite erupted volume until a sheet ‘a’a, pahoehoe, transitional, blocky, and infla-
with a critical thickness defined by the lava tionary), can build a range of flow field architec-
yield strength is attained. For a cooling-limited tures spanning simple, single-unit lava flows to
model, heat losses due to radiation, convection, compound lava flow fields comprising hundreds
and conduction are estimated to calculate core to hundreds of thousands of individual lava flow
cooling and crystallization rates. These can be units (e.g., Walker, 1972; Kilburn and Lopes,
used to estimate the lava rheological properties 1988; Mattox et al., 1993). Each system can be
(viscosity, yield strength) of the lava. Rheological channel-, tube-, or channel-and-tube-fed, with
properties can be used to estimate flow thick- the feeder system often bifurcating to result in
ness through a yield strength model, as well multiple pathways which increase in number
as velocity using the Jeffreys or Navier–Stokes with distance from the vent (e.g., Peterson etâ•–al.,
equations. Such flow simulation models rely, 1994; Calvari and Pinkerton, 1998). It is the
in turn, on well-constrained models that define challenge of the numerical modeler to simulate
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
and link the complex interplay between topog- allowing generation of simulations to identify
raphy, rheology, heat loss, and flow dynamics likely inundation zones during ongoing erup-
that lead to the emplacement of a specific lava tions (e.g., Tedesco et al., 2007) or assessment of
flow system of given morphology, architecture, likely lava flow hazard zones for future effusive
thickness, length, and width. By fitting lava flow eruptions (e.g., Rowland et al., 2005), as well as
models to real-world systems we can achieve two evaluation of downflow changes in dynamic,
aims: (i) to understand the complex rheological, thermal, and rheological properties of a spread-
thermal, and dynamic processes involved in ing lava. This chapter focuses on the mechanics
lava flow emplacement, and which give rise of Type II models, first reviewing the underlying
to specific flow emplacement styles, lava flow fluid dynamic, rheological, and heat loss princi-
types architectures, morphologies, and dimen- ples, many of which draw on the results of Type
sions; and (ii) to more accurately simulate lava I modeling, before examining the way that Type
flow emplacement events. II models function.
Consequently, lava flow numerical models
can be generalized into two types: (I) those that
improve our understanding of emplacement 5.2╇ ╇Lava flow dynamics:
dynamics and lava flow morphology, and (II)
those designed to simulate lava flow emplace- fundamental principles and
ment. Type I models (reviewed in online definitions
Supplement 5A€ – see end of chapter) can be
split into theoretical or laboratory-based, with The principles governing lava flow dynam-
laboratory-based studies commonly using kao- ics were laid down by Hulme (1974) and later
lin or “wax” as an analog for lava in scaled, detailed by Chester et al. (1985). We begin by
tank-based, simulations. Both theoretical and defining two fundamental terms: shear stress, τ
laboratory approaches typically attempt to par- (Pa or N m–2), and strain rate ε̇ (s–1). Shear stress
ameterize links between lava flow heat loss, is the force per unit area acting on a fluid and
cooling and rheology, and the resulting flow depends on flow thickness, h, and density, ρ, as
dynamics, dimensions, and morphology. Many well as gravity, g, and the slope, θ, over which
laboratory models relate flow dimensions or the fluid, in this case lava, is moving:
morphology to effusion rates by applying laws
τ = hâ•›ρg sin θ. (5.1)
governing heat transfer to explain observed rela-
tionships (e.g., Fink and Griffiths, 1992; Blake (All notation is given in Section 5.7.) Strain rate
and Bruno, 2000). Theoretical approaches can is the rate of deformation experienced by a fluid
include empirical attempts to link flow dimen- when a load stress is applied, and can be defined
sional properties, such as length and planform by the flow’s velocity gradient (i.e., ε̇ = du/dz).
area, to volumetric eruption rates. Such models We next distinguish between Newtonian
are reviewed by Kilburn et al. (1995) and Harris and Bingham fluids. A Newtonian fluid will flow
and Rowland (2009) but are not considered in (deform) when an infinitesimally small force
Supplement 5A. (shear stress) is applied. In contrast, a Bingham
Type II models use fluid dynamic and heat fluid will only flow once sufficient shear stress
transfer principles to predict flow emplacement has been applied. If a Bingham fluid is not suf-
by simulating likely flow direction and extent, as ficiently thick, shear stresses will not be high
summarized in online Supplement 5B (see end enough to cause a deformation response, so
of chapter). Type II studies commonly use find- that strain rate will be zero. Upon attaining a
ings from Type I modeling to evaluate cooling- critical thickness, h0, shear stresses in Bingham
induced changes in flow rheology, and hence fluids become sufficient to cause deformation or
to determine how flow forward motion and/or flow. For a Bingham fluid on an inclined plane,
spreading varies in space and time. As such, Type this point is defined by the yield strength of the
II models are often used for hazard assessment, fluid:
c
vis
Ps
Yield strength can also as be written in terms of 5.2.1â•…Lava flow dynamics: velocity
the height-to-width ratio that will be attained by treatments
a fluid spreading under the influence of gravity The fluid dynamics of viscous fluids are des�
on a flat surface (Blake, 1990). For a given vol- cribed extensively in engineering text books
ume, such a fluid will spread until it attains a (e.g., White, 2006), as well as for geological appli-
critical thickness, h0, and width, w, to produce cations by Furbish (1997), and many numerical
a stationary pile of stationary (non-deforming) treatments for viscous flow appear in the fluid
fluid: mechanics literature. Mei and Yuhi (2001), for
example, give one of the most complete solu-
τ0 = gâ•›ρh0 / w. (5.2b) tions for three-dimensional flow of a Bingham
fluid in a wide, shallow channel (see Balmforth
Bingham fluids will only spread as long as et al. (2006) for review).
τ > τ0; once τ ≤ τ0, the flow will have reached For lava flows, the relation of Jeffreys (1925)
its minimum thickness h0, and spreading will has become widely used. This provides a rela-
stop. In contrast, Newtonian fluids have no tion between fluid viscosity, η, and mean vel-
yield strength and the fluid can spread to form ocity, u, for gravity-driven Newtonian flow in
an infinitesimally thin sheet. one dimension:
h 2 g ρsin θ wh3 g ρ sin θ 3 τ 0 1 τ 0
3
η= , (5.3a) Er = 1 − + . (5.5)
Bu 3η 2 τb 2 τb
in which h is flow depth, θ is the slope of the For flow in two dimensions, the equation of
underlying surface, and B is a constant with a motion from Navier (1823) and Stokes (1845)
value of 8 for semi-circular channels and 3 for can be applied. For steady state conditions, the
channels that are wider than they are deep. Navier–Stokes equations can be used to assess
Equation (5.3a) was originally applied to estimate the balance of forces incident upon the flow.
lava flow viscosity using velocity and slope data As shown by Miyamoto and Sasaki (1997), the
(e.g., Walker, 1967), but simple re-arrangement Navier–Stokes equations can be used to con-
allows calculation of mean velocity if viscosity sider motion in the vertical (z) and downflow (x)
is known: dimensions:
h 2 g ρsin θ ∂2u ∂ h (5.6)
u= . (5.3b) η = ρ g cos θ + ρ g sin θ.
Bη ∂z2 ∂x
In effect, Eq. (5.3b) considers the balance This describes a scenario where viscous forces
between gravitational forces (numerator) and (left-hand side of Eq. (5.6)) are balanced by grav-
resistive forces (denominator) on the flow. ity and pressure variations induced by changes in
This equation is, however, only appropriate flow thickness, h (right-hand side of Eq. (5.6)). If
for flow of a Newtonian fluid in a channel (and the lava undergoes little or no variation in thick-
so is not appropriate for calculating flow-front ness, then the pressure term can be neglected so
velocities). that Eq. (5.6) can be written solely in terms of
Moore (1987) provides a treatment for a viscous and gravitational forces (Dragoni et al.,
Bingham fluid that takes into account yield 1986):
strength, where for flow in a semi-circular
channel: ∂2 u (5.7)
η − ρ g sin θ = 0.
∂z 2
h 2 g ρ sin θ
4
4 τ0 1 τ0 Integration of this equation allows the vertical
u= 1 − + , (5.4a)
8η 3 τ b 3 τ b and horizontal profiles of flow velocity within
and across a flow to be calculated (Fig. 5.2).
and for a channel that is wider than it is deep: Alternatively, the Navier–Stokes equations
can be used to describe motion in both the down-
and cross-flow (x and y) directions, described by
h 2 g ρsin θ 3 τ 0 1 τ 0
3
(a)
ρ = density (2030 kg m–3) θ = slope (5°)
u(z) = ρ g sin(θ) (t 2-z 2) DRE ρ of 2600 kg m–3 corrected for t = channel depth or radius (2 m)
4µ 22% vesicles z = flow depth or distance from bank
g = acceleration due to gravity (9.8 m–2) µ = viscosity (3000 Pa s)
(b)
0.0 (c) 0.6
0.5
0.4
Depth below surface, z (m)
mean
1.2
0.2
1.6
0.1
2.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 2.0 1.0 0.0 1.0 2.0
Velocity, u(z) (m s–1) Distance from channel center line, z (m)
Figure 5.2╇╇Velocity profile equation obtained from units of one or more compounds (monomers).
integration of Eq. (5.7) and parameters used to derive the The presence of polymers leads to highly inter-
variation in flow velocity with depth and cross-channel connected networks of molecules that pose
distance as given in (b) and (c). Input parameters apply to a internal resistance to flow. Because the strong-
channel observed as active on Etna during 2001 by Bailey
est bonds in silicate liquids are the Si-O-Si
et al. (2006). Calculated maximum velocities compare well
bridging bonds, the internal resistance of the
with values of ~0.6 m s–1 measured during brim-full flow.
fluid to flow (i.e., viscosity) increases with
silica content. However, network-modifying
Inspection of these equations reveals that oxides and water serve to break up the bridg-
defining and linking three sets of variables is ing bonds, thereby depolymerizing the fluid
crucial: (i) flow or channel dimensions (w, h), and reducing flow viscosity. Temperature also
(ii) slope (θ), and (iii) rheology (η, τ0). If we can influences viscosity: increased temperature
understand the variation in these parameters in provides increased energy, thereby weakening
space and time, and numerically link them to the strength of the bonds in the melt and low-
the resulting flow dynamics, we can arrive at an ering the viscosity. Below liquidus tempera-
effective Type II lava flow model. Of these, rhe- tures, crystallization also plays a role, whereby
ology is the most complicated variable to set, the presence of the rigid particles (crystals)
given that it depends on lava composition, tem- impedes flow. The presence of bubbles in the
perature, crystallinity, and vesicularity. Thus, melt also influences the internal resistance
numerical relationships between, for example, of magma. Viscosity relations for fluids, and
viscosity and temperature need to be defined. fluids containing particles, are considered in
numerous rheological texts, such as Larson
5.2.2â•…Controls on viscosity and yield (1999), and discussed in the context of conduit
strength ascent models in Chapter 4. Here we provide a
Lava is a polymer, i.e., a compound whose mol- brief review of treatments relevant to silicate
ecule is formed from a number of repeated lava flow models.
103 106
2
10
104 Spherical bubbles
0.4 wt.%
103
10
102
2.0 wt.%
10 Sheared bubbles
1 µ
3.4 wt.%
1
0 0
1000 1050 1100 1150 1200 0 0.2 0.4 0.6 0.8 1.0
Temperature, T (°C) Bubble fraction, ω
105 105
Case: 1
Mixture viscosity, µm (Pa s)
φmax = 0.6 2
104 104 3
φmax = 0.7
103 103
µ µ (with 35 % bubbles)
102 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 0.6
Crystal fraction, φ Crystal fraction, φ
Figure 5.3╇╇ (a) Melt viscosity for Etna’s alkali basalt as a function of temperature following the relation of Giordano and
Dingwell (2003):
log10η = –4.643 + [5812.44€– 427.04 × H2O] / [T(K)€– 499.31 + 28.74 × ln(H2O)])
The relation is given for the range of water contents associated with Etna’s undegassed (2.0–3.4 wt. %) and degassed magma
(0.1–0.4 wt. %). (b) Relation between mixture viscosity and crystallinity defined by Eq. (5.11) using ϕmax of 0.5 and 0.7. (c) Relation
between mixture viscosity and bubble content defined by Eq. (5.12a) (spherical bubble relation) and Eq. (5.12b) (sheared bubble
relation). (d) Three-phase mixture viscosity as a function of crystallinity defined using Eq. (5.13a) (Case 1), Eq. (5.13b) (Case 2),
and Eq. (5.13c) (Case 3) using a bubble volume fraction of 0.35. Each mixture viscosity relation is set using fluid viscosity (η) of
425 Pa s. This is calculated using the relation of Giordano and Dingwell (2003) for a degassed lava (0.1 wt. % H2O) erupted at a
temperature of 1086°C and defines the y-axis offset in (b), (c), and (d). Black bar in (b) and (d) indicates crystal content (45%) and
viscosity range (9400 ± 1500 Pa s) measured for an Etnean lava by Pinkerton and Sparks (1978).
on the basis of crystal concentration, size, and This heat is then lost from the surface by radi-
shape, with Ryerson et al. (1988) showing that ation and convection. Radiation from a lava sur-
data for picritic basalt from Kilauea Iki are best face at temperature Tsurf can be assessed using
Â�fitted with a power law expression whereby the Stefan–Boltzmann law. This describes the
τ0 = 6500ϕ2.85. energy flux (qrad in W m–2) radiated directly from
The challenge with lava is that, unlike sim- the surface of a gray-body, i.e., a body with an
ple Bingham fluids, high temperature gradients emissivity ε < 1:
mean that severe rheological gradients occur
q rad = εσ (Tsurf
4
− Ta4 ), (5.15)
vertically within a lava flow, such that no single
rheology applies. Lateral spreading, for example,
will be influenced by the development of a cool, in which σ is the Stefan–Boltzmann constant
high viscosity, viscoelastic skin. This means that (5.67 × 10–8 W m–2 K-4) and Ta is the tempera-
the fluid core will be confined within a strong ture of the environment into which radiation
outer shell (Hon et al., 1994). As a result, one is being emitted. Multiplying by the area of the
modeling direction has been to examine the radiating surface (A) gives the total radiant heat
ability of such rheologically defined outer shells loss, Qrad, in watts. The first question, is how to
to contain a steadily pressurizing or inflating determine Tsurf ?
core (e.g., Iverson, 1990) and to examine the Crisp and Baloga (1990) suggested that the
role of this outer shell in influencing flow front surface of an active lava is best modeled by a
break outs (e.g., Dragoni and Tallarico, 1996). two-component thermal surface in which a
chilled crust at temperature Tcrust occupies frac-
tion f of the surface, with cracks exposing the
molten core at Tcore occupying the remainder. As
5.3╇ ╇Lava flow heat budget a result, Tsurf is best described by the effective
and cooling radiation temperature (Te):
1/ 4 (5.16)
Because temperature, and cooling-induced Te = fTcrust
4
+ (1 − f )Tcore
4
.
changes in crystallinity, play central roles in
determining flow rheology, understanding the In a stable case (when there is no destruction
flow heat budget and cooling rates is an import- and renewal of the crust, so that it simply thick-
ant step in achieving a Type II model. This ens with time), the cooler of these two thermal
involves the derivation of a complete lava flow components can be determined on the basis of
heat budget that defines the balance between time since exposure (t). For Hawaiian pahoe-
heat sinks and heat sources. While Keszthelyi hoe, measurements by Hon et al. (1994) showed
(1995) provides a complete heat loss model that surface temperature declined logarithmic-
for tube-confined lava, general texts such as ally with time (Tcrust = –140 log(t) + 303) as the
Holman (1992) detail the principles of heat surface crust thickened with the square root of
transfer. We here consider the heat budget time (D1070 = 0.0779√t). Here, the 1070 subscript
for channel-confined lava and/or dispersed shows that this relation is for the depth of the
(‘a’a, pahoehoe and blocky) lava flows whose 1070â•›°C isotherm, which is used to define the
surfaces are exposed to the atmosphere. base of the visco-elastic surface crust. On the
For an active channel-fed ‘a’a, pahoehoe, face of it, the hotter of the two components
or blocky lava flow, the main heat sinks will appears easy to set, being essentially equal to the
be radiation and convection from the flow sur- lava core temperature. In reality, the maximum
face (e.g., Keszthelyi et al., 2003), with conduc- temperature of the flow core is rarely exposed
tion causing heat loss through the flow levées at the surface. Instead, cracks penetrate a cooler
(e.g., Quareni et al., 2004) and base (e.g., Fagents viscoelastic layer that forms just below the brit-
and Greeley, 2001). Heat is conducted across a tle crust and which, for Hawaiian pahoehoe, has
surface crust that forms over the molten core. a maximum temperature of 1070â•›°C, as opposed
to a core temperature of 1140–1150â•›°C (Hon et al., in terms of a heated plate exposed to ambient
1994), meaning that maximum surface tem- room air without an external source of motion
peratures measured at Hawaiian pahoehoe are (i.e., no wind). Heating of the air from beneath
typically ~1000â•›°C. Likewise, maximum surface causes the air to expand. The subsequent reduc-
temperatures measured at a lava channel on Mt. tion in density causes the heated air to become
Etna (Italy) within 70 m of the vent were typic- buoyant so that it rises, carrying heat with it and
ally 730–1040â•›°C (mean of 880â•›°C) compared with away from the surface. In contrast, forced con-
a core temperature of 1065â•›°C (Harris et al., 2005; vection is the cooling effect experienced when
Bailey et al., 2006). Thus, use of a value for Tcore cool air blows over a hot surface. Both can be
that is offset from the true core temperature may described in terms of the convective heat trans-
yield more realistic values of Te. In addition, more fer coefficient, hc:
than one thermal component may be present on
qconv = hc (Tsurf – Ta), (5.17)
the flow surface (Figs. 5.4(b), 5.4(c)). Wright and
Flynn (2003) found that active pahoehoe surfaces where qconv is the convective heat flux (in W m–2),
were best modeled using between five and seven and multiplying by the area, A, of the hot sur-
thermal components at temperatures ranging face gives the total convective heat loss, Qconv, in
between 200 and 1000â•›°C. In comparison, ther- watts. The convective heat transfer coefficient,
mal data from active lava channels indicate that hc, can be calculated theoretically (as in Exercise
a four thermal component model may be used to 5.2; online resources listed at end of chapter),
describe such surfaces (see Exercise 5.1; online but differs for free and forced convection and
resources listed at end of chapter). varies with wind speed. However, values of
Convective heat loss from a flow surface between 5 and 150 W m–2 K–1 appear appropri-
involves heat transfer to an overlying gas or ate for subaerial active lavas (Keszthelyi and
fluid. There are two cases. Free convection occurs Denlinger, 1996; Neri, 1998; Keszthelyi et al.,
in still conditions and can be conceptualized 2003).
be estimated to allow calculation of Qrad and effect this has on inhibiting further flow advance
Qconv, and (iii) a levée and basal conduction heat through modifying the flow core rheology. In
loss model is applied, then the core cooling rate volume-limited cases, the flow stops once the
can be calculated given reasonable assumptions available volume has been used up, irrespective
for ρ, cp, and crystallization rates. Exercises 5.3 of the cooling and/or rheological conditions. All
and 5.4 (see online resources listed at end of Type II flow models also require a topographic
chapter) illustrate this procedure. model to determine the likely path that a fluid
will follow. We thus begin with an assessment
5.3.1â•… Crystallinity of flow path projection routines.
Viscosity, yield strength, and cooling rates also
depend on crystallinity, which will increase 5.4.1â•… Flow paths
with decreasing flow core temperature. Crystal The simplest method to generate a flow path is
content can be expressed as a function of time to identify the course that water draining down
using the relationship defined by Crisp and a slope would follow (e.g., Fig. 5.5). Application
Baloga (1994): of such flow path models are commonplace in
hydrology and are available as part of many
ϕ(t) = ϕmax [1–exp(–t/tϕmax), (5.23)
Geographical Information Systems (GIS). Many
in which tϕmax is the time needed to reach the routines are based on the deterministic eight
maximum crystallinity (ϕmax). Alternatively, crys- neighbors (D8) method (Mark, 1984; O’Callaghan
tallization rates can be obtained as a function of and Mark, 1984). This routine finds the path of
distance using the cooling rate per unit distance steepest descent across a gridded topography or
(Harris and Rowland, 2001): Digital Elevation Model (DEM). To do this, the gra-
dient between a pixel and all eight of its neigh-
dφ dT dφ
= , (5.24) bors is calculated. The flow line is then projected
dx dx dT from the host pixel in the direction of most nega-
tive gradient. This process is continued from
in which dϕ/dT is volume fraction of crystals pixel to pixel to define the flow line. The basic D8
grown (dϕ) over a given temperature drop (dT). algorithm has problems with long planar slopes,
as well as flat areas and pits, i.e., cases where the
host pixel will be lower than, or the same ele-
5.4╇ ╇Type II modeling vation as, all surrounding pixels (Fairfield and
Leymarie, 1991; Turcotte et al., 2001; Jones, 2002).
The aim of a Type I model is to draw links Pits may either be natural or due to noise in the
between any two of the following: (i) heat loss DEM. To solve these problems, a number of mod-
and core cooling, (ii) flow rheology, (iii) flow ifications have been proposed to the D8 model
dynamics and dimensions, and/or (iv) lava sur- including use of a stochastic function to solve
face morphology. The aim of the Type II model flow path problems over planar slopes (Fairfield
is to draw these findings together to generate and Leymarie, 1991), use of maximum gradients
a complete lava flow emplacement model. The calculated using eight triangular facets centered
main attempts to achieve this are summarized on the host pixel to improve flow line predic-
in Supplement 5B (online resource€– see end of tion (Tarboton, 1997), smoothing to remove pits
chapter). (Hutchinson, 1989), or allowing the flow line to
The foundation of Type II modeling is leave a pit through the lowest pixel on the pit
either (a) volume-limited flow emplacement, rim (Fairfield and Leymarie, 1991; Tribe, 1992).
or (b) cooling-limited flow emplacement. A In the latter case, once the pit outlet has been
cooling-limited flow is one whereby supply (lava found, a flow line is projected across the pit from
effusion) is maintained until the full flow extent the entrance to exit point. This creates a flat area
is realized. That is, the limits of the flow unit are across the pit. To solve for problems in flat areas,
defined by the processes of lava cooling and the Martz and Garbrecht (1998) proposed a solution
Figure 5.6╇╇ (a) Maximum gradient path (yellow line) from DEM quality
Kilauea’s July 1974 vent location to the DEM edge (coast). In all flow path prediction algorithms, DEM
(b) Stochastic flow path model from Kilauea’s July 1974 vent quality is an issue; the DEM needs to be as
location to the coast. Model applies DEM noise of 2 m and smooth and noise-free as possible, but without
has been run for 20 iterations. (c) Twenty-iteration flow line losing topographic detail that may effect the
runs with 2 m of DEM noise for Kilauea’s July and December flow path. On a fine scale, flows moving over
1974 flows. Flow lines have been cut at the flow toe in each
existing pahoehoe and ‘a’a lavas will be influ-
case to allow an assessment of the fit with reality, as revealed
by the black flow areas in this Thematic Mapper (TM) color
enced by inflation features, such as tumuli, or
composite. Simulations have been projected onto a near true by channels, levées, and flow margins. It is not
color (vegetation = bright green, lava = black) TM image of uncommon, for example, for a flow to follow
Kilauea. See color plates section. a pre-existing flow margin (which represents a
local topographic high), or even re-use a drained
lava channel or tube (which represents a local
from points located around the rim of a catch- topographic low). Also, care needs to be taken
ment (i.e., along the watershed) will converge to to remove the influence of trees from eleva-
exit the catchment at a single point, as shown tion data. If elevations are obtained from the
in Figure 5.5. Thus, lava flows from multiple tree-canopy, DEM-projected flow paths may fol-
sources along a rift zone may be predicted to low false boundaries induced by the apparent
enter the ocean at the same place. Such an scarp caused by a forest edge bordering a lava
approach was used by Guest and Murray (1979) flow field or road. The good fit for the modeled
in assessing lava flow hazard on Etna, where all area of Kilauea’s December 1974 flow in Figure
possible flow paths were generated by project- 5.6(c), for example, is due to the simulation fill-
ing downhill lines from all likely vent zones. ing a false depression in the DEM caused by
These converged to define a finite number of the emplacement of the real lava flow, which
highest-order stream points at the base of Etna’s destroyed the trees in this otherwise forested
cone at which flows would likely arrive. area. Such false depressions due to tree clearance
Sasaki (1997), for example, yield strength, vis- are providing increasingly better fits with the
cosity, thickness, and slope are varied from cell characteristics (e.g., effusion rate, rheological
to cell as the lava spreads across the topographic conditions, cooling, velocity, flow length, inun-
grid obeying the equation of motion given in dation area, and emplacement duration) of
Eq. (5.6). The MAGFLOW model is instead built the real-world flows against which they are
on the Cellular Nonlinear Network model of compared. However, one issue that has not
Del Negro et al. (2005), which applies Eq. (5.8) been explored so far in this chapter is simula-
to also take into account cross-flow velocity. tion run time. Cellular automata models are
However, it too uses a cooling model to modify computationally complex, but allow for more
yield strength, viscosity, and flow thickness as complex flow geometries to be modeled than
the lava spreads from cell to cell across a topo- one-dimensional models, allowing flow spread-
graphic grid. ing in two dimensions to be simulated. However,
The SCIARA cellular automata model is dif- the complexity of the models makes run times
ferent in that it is not based on the Navier– quite long. Even the two-dimensional model of
Stokes equations (Crisci et al., 2004). Instead, Miyamoto and Sasaki (1997), for example, has
the flow in and out of the cells is determined run times of up to two days when executing a
by hydrostatic pressure gradients due to dif- 40-day simulation on a 3 GHz computer (~1.5
ferences in lava thickness within surrounding hours for each day of simulation). The FLOWGO
cells, and the rheology of lava in the cell (Barca model of Harris and Rowland (2001) is computa-
et al., 1993). Rather like FLOWFRONT, and all tionally much less complex, allowing run times
other CA models, it relies on an assessment of for a single flow path of less than one second
the critical thickness of lava that must remain (e.g, Fig. 5.6(a)). Running multiple flow paths in
in a given cell. Unlike FLOWFRONT, where a a stochastic simulation increases run times but
fixed yield strength (that does not vary down- only to 30 seconds for a 20-iteration run (e.g.,
flow) is used to define this critical thickness, Fig. 5.6(b).
in SCIARA the critical thickness is set using If models are to be used for real-time hazard
an “adherence parameter,” ϖ. This defines assessment, for example, in the case of a rap-
the thickness of lava that cannot flow out of idly advancing flow on a heavily populated vol-
a cell due to rheological resistance and varies cano, run times need to be low so that results
with temperature following the form of the are available in a timely fashion. If a flow is
expected temperature-dependent relationship moving towards a town at 10 km h–1, even if
for yield strength (ϖ = ae–bT). Thus, like all Type the town is 25 km distant from the source vent,
IIb models, SCIARA is driven by a heat budget there is little time available to run a simulation
based cooling model that is used to assess the to identify vulnerable infrastructure and zones
temperature of lava in any given cell, and hence in need of immediate evacuation. This was
its rheological conditions. the case of Mauna Loa’s 1950 eruption, which
began at ~10:15 pm on 1 June 1950. Within 2
hours and 15 minutes the flow had extended
23 km to enter the town of Hookena, where it
5.5╇ ╇ Discussion destroyed the main highway, post office, sev-
eral houses and a gas filling station in short
Lava flow emplacement models are becoming order (Finch and Macdonald, 1950). In such a
increasingly sophisticated. Mostly this is because scenario, there are less than 2.5 hours available
of our ability to understand and parameterize to inform the decision-making process. Civil
the interplay among various flow emplacement Protection authorities responding to such an
dynamics has improved immensely in the last event today would require information within
three decades, as witnessed by the weight of minutes. Given reliable lava flow simulation
literature that comprises online Supplement models, one aim must now be to reduce run
5A. As a result, flow emplacement simulations times to a minimum.
melt viscosity: A treatment for met aluminous, volcanoes in Japan. In Lava Flows and Domes, ed.
peraluminous and peralkaline liquids. Chemical J. Fink. Berlin: Springer, pp. 184–207.
Geology, 229, 42–56. Iverson. R. M. (1990). Lava domes modeled as brittle
Gregory, K. J. and Walling, D. E. (1973). Drainage Basin shells that enclose pressurized magma, with
Form and Process. London: Edward Arnold. application to Mount St. Helens. In Lava Flows and
Guest, J. E. and Murray, J. B. (1979). An analysis of Domes, ed. J. Fink. Berlin: Springer, pp. 48–69.
hazard from Mount Etna volcano. Journal of the Jeffreys, H. (1925). The flow of water in an inclined
Geological Society of London, 136, 347–354. channel of rectangular section. Philosophical
Harris, A. J. L. and Allen III, J. S. (2008). One-, two- and Magazine, 49, 793–807.
three-phase viscosity treatments for basaltic lava Jones, R. (2002). Algorithms for using a DEM
flows. Journal of Geophysical Research, 113, B09212, for mapping catchment areas of stream
doi:10.1029/2007JB005035. sediment samples. Computers and Geosciences, 28,
Harris, A. and Rowland, S. (2001). FLOWGO: a 1051–1060.
kinematic thermo-rheological model for lava Kauahikaua, J., Margriter, S., Lockwood, J. and
flowing in a channel. Bulletin of Volcanology, 63, 20–44. Trusdell, F. (1995). Applications of GIS to the
Harris, A. J. L. and Rowland, S. K. (2009). Effusion rate estimation of lava flow hazards on Mauna Loa
controls on lava flow length and the role of heat Volcano, Hawai’i. In Mauna Loa Revealed: Structure,
loss: A review. In Studies in Volcanology: The Legacy of Composition, History and Hazards, ed. J. M. Rhodes
George Walker, ed. T. Thordarson, S. Self, and J. P. Lockwood, American Geophysical Union,
G. Larsen, S. K. Rowland and A. Hoskuldsson, Geophysical Monograph 92, pp. 315–325.
Special Publications of IAVCEI, 2, 33–51. Kerr, R. (2001). Thermal erosion by laminar lava
Harris, A., Bailey, J., Calvari, S. and Dehn, J. (2005). flows. Journal of Geophysical Research, 106(B11),
Heat loss measured at a lava channel and its 26,453–26,465.
implications for down-channel cooling and Keszthelyi, L. (1995). A preliminary thermal budget
rheology. Geological Society of America Special Paper for lava tubes on the Earth and planets. Journal of
396, 125–146. Geophysical Research, 100(B10), 20411–20420.
Hess, K.-U. and Dingwell, D. B. (1996). Viscosities Keszthelyi, L. and Denlinger, R. (1996). The initial
of hydrous leucogranitic melts: A non-Arrhenian cooling of pahoehoe flow lobes. Bulletin of
model. American Minerologist, 81, 1297–1300. Volcanology, 58, 5–18.
Hidaka, M., Umino, S. and Fujita, E. (2005). VTFS Keszthelyi, L. and Self, S. (1998). Some physical
project: Development of the lava flow simulation requirements for the emplacement of long basaltic
code LavaSIM with a model for three-dimensional lava flows. Journal of Geophysical Research, 103(B11),
convection, spreading, and solidification. 27â•›447–27â•›464.
Geochemistry, Geophysics, Geosystems, 6, Q07008, Keszthelyi L., Harris, A. and Dehn, J. (2003).
doi:10.1029/2004GC000869. Observations of the effect of wind on the cooling
Holman, J. P. (1992). Heat Transfer. London: of active lava flows. Geophysical Research Letters,
McGraw-Hill. 30(19), doi:10.1029/2003GL017994.
Hon, K., Kauahikaua, J., Denlinger, R. and Kilburn, C. R. J. and Lopes, R. M. C. (1988). The
Mackay, K. (1994). Emplacement and inflation growth of aa lava fields on Mount Etna, Sicily.
of pahoehoe sheet flows: Observations and Journal of Geophysical Research, 93, 14â•›759–14â•›772.
measurements of active lava flows on Kilauea Kilburn, C. R. J., Pinkerton, H. and Wilson, L.
Volcano, Hawaii. Geological Society of America (1995). Forecasting the behaviour of lava flows. In
Bulletin, 106, 351–370. Monitoring Active Volcanoes, ed. W. McGuire,
Hulme, G. (1974). The interpretation of lava flow C. R. J. Kilburn and J. B. Murray. London: UCL
morphology. Geophysical Journal of the Royal Press, pp. 346–368.
Astronomical Society, 39, 361–383. Larson, R. G. (1999). The Structure and Rheology of
Hutchinson, M. F. (1989). A new procedure for Complex Fluids. New York: Oxford University
gridding elevation and stream line data with Press.
automatic removal of spurious pits. Journal of Llewellin, E. W. and Manga, M. (2005). Bubble
Hydrology, 106, 211–232. suspension rheology and implications for conduit
Ishihara, K., Iguchi, M. and Kamo, K. (1990). flow. Journal of Volcanology and Geothermal Research,
Numerical simulation of lava flows on some 143, 205–217.
Lipman, P. W. and Banks, N. G. (1987). Aa flow Pinkerton, H. and Sparks, R. S. J. (1978). Field
dynamics, Mauna Loa. U.S. Geological Survey measurements of the rheology of lava. Nature, 276,
Professional Paper, 1350, 1527–1567. 383–385.
Manga, M., Castro, J., Cashman, K. V. and Pinkerton, H. and Stevenson, R. J. (1992). Methods of
Loewenberg, M. (1998). Rheology of bubble-bearing determining the rheological properties of magmas
magmas. Journal of Volcanology and Geothermal at sub-liquidus temperatures. Journal of Volcanology
Research, 87, 15–28. and Geothermal Research, 53, 47–66.
Mano, J. F. and Pereira, E. (2004). Data analysis with Quareni, F., Tallarico, A. and Dragoni, M. (2004).
the Vogel-Fulcher-Tammann-Hesse equation. Modeling of the steady-state temperature field in
Journal of Physical Chemistry, 108, 10â•›824–10â•›833. lava flow levees. Journal of Volcanology and Geothermal
Mark, D. M. (1984). Automated detection of Research, 132, 241–251.
drainage networks from digital elevation models. Rowland, S., Garbeil, H. and Harris, A. (2005).
Cartographica, 21, 168–178. Lengths and hazards from channel-fed lava flows
Marsh, B. D. (1981). On the crystallinity, probability on Mauna Loa, Hawai’i, determined from thermal
of occurrence, and rheology of lava and magma. and downslope modeling with FLOWGO. Bulletin of
Contributions to Mineralogy and Petrology, 78, 85–98. Volcanology, 67, 634–647.
Martz, L. W. and Garbrecht, J. (1998). The treatment Russell, J. K., Giordano, D. and Dingwell, D. B.
of flat areas and depressions in automated (2003). High-temperature limits on viscosity of
drainage analysis of raster digital elevation non-Arrhenian silicate melts. American Minerologist,
models. Hydrological Processes, 12, 843–855. 88, 1390–1394.
Mattox, T. N., Heliker, C., Kauahikaua, J. and Hon, Ryerson, F. J., Weed, H. C. and Piwinskii, A. J. (1988).
K. (1993). Development of the 1990 Kalapana Rheology of subliquidus magmas 1. Picritic
flow field, Kilauea Volcano, Hawaii. Bulletin of compositions. Journal of Geophysical Research, 93(B4),
Volcanology, 55, 407–413 3421–3436.
Mei, C. and Yuhi, M. (2001). Slow flow of a Bingham Shaw, H. R. (1972). Viscosities of magmatic silicate
fluid in a shallow channel of finite width. Journal of liquids: an empirical method of prediction.
Fluid Mechanics, 431, 135–159. American Journal of Science, 272, 870–893.
Miyamoto, H. and Sasaki, S. (1997). Simulation of Sonder, I., Zimanowski, B. and Buttner, R.
lava flows by an improved cellular automata (2006). Non-Newtonian viscosity of basaltic
method. Computers and Geosciences, 23, 283–292. magma. Geophysical Research Letters, 33, L02303,
Moore, H. J. (1987). Preliminary estimates of the doi:10.1029/2005GL024240.
rheological properties of 1984 Mauna Loa lava. U.S. Stevens, N. F., Wadge, G. and Murray, J. B. (1999).
Geological Survey Professional Paper, 1350, 1569–1588. Lava flow volume and morphology from digitised
Navier, C. L. M. H. (1823). Mémoire sur les lois du contour maps: a case study at Mount Etna, Sicily.
mouvement des fluids. Memoirs of the Academy Royale Geomorphology, 28, 251–261.
Paris, 6, 389–416. Stokes, G. G. (1845). On the theories of internal
Neri, A. (1998). A local heat transfer analysis of friction of fluids in motion. Transactions of the
lava cooling in the atmosphere: application to Cambridge Philosophical Society, 8, 287–305.
thermal diffusion-dominated lava flows. Journal of Tarboton, D. G. (1997). A new method for the
Volcanology and Geothermal Research, 81, 215–243. determination of flow determinations and upslope
O’Callaghan, J. F. and Mark, D. M. (1984) The areas in grid digital elevation models. Water
extraction of drainage networks from digital Resources Research, 33(2), 309–319.
elevation data. Computer Vision, Graphics and Image Tedesco, D., Badiali, L., Boschi, E. et al. (2007).
Processing, 28, 323–344. Cooperation of Congo volcanic and environmental
Peterson, D. W., Holcomb, R. T., Tilling, R. I. and R. risks. Eos, Transactions of the American Geophysical
L. Christiansen (1994). Development of lava tubes Union, 88(16), 177–181.
in the light of observations at Mauna Ulu, Kilauea Tribe, A. (1992). Automated recognition of valley
volcano, Hawaii. Bulletin of Volcanology, 56, 343–360. lines and drainage networks from grid digital
Phan-Thien, N. and Pham, D. C. (1997). Differential elevation models: a review and a new method.
multiphase models for polydispersed suspensions Journal of Hydrology, 139, 263–293.
and particulate solids. Journal of Non-Newtonian Fluid Turcotte, D. L. and Schubert, G. (2002). Geodynamics.
Mechanics, 72, 305–318. Cambridge: Cambridge University Press, 456 pp.
Turcotte, R., Fortin, J.-P., Rousseau, A. N., Whittington, A., Richet, P. and Holtz, F. (2000). Water
Massicotte, S. and Villeneuve, J.-P. (2001). and the viscosity of depolymerized aluminosilicate
Determination of the drainage structure of a melts. Geochimica et Cosmochimica Acta, 64(21),
watershed using a digital elevation model and a 3725–3736.
digital river and lake network. Journal of Hydrology, Wright, R. and Flynn, L. P. (2003). On the retrieval
240, 225–242. of lava-flow surface temperatures from infrared
Vicari, A., Herault, A., del Negro, C. et al. (2007). satellite data. Geology, 31(10), 893–896.
Modeling of the 2001 lava flow at Etna volcano Young, P. and Wadge, G. (1990). FLOWFRONT:
by a Cellular Automata approach. Environmental Simulation of a lava flow. Computers and Geosciences,
Modelling and Software, 22, 1464–1471. 16, 1171–1191.
Walker, G. P. L. (1967). Thickness and viscosity of
Etnean lavas. Nature, 213, 484–485. Online resources available at www.cambridge.
Walker, G. P. L. (1972). Compound and simple lava org/fagents
flows and flood basalts. Bulletin of Volcanology, 35,
579–590. • Supplement 5A: Review of Type I models
White, F. M. (2006). Viscous Fluid Flow. New York: • Supplement 5B: Review of Type II models
McGraw-Hill, 629 pp. • Exercises: Construction of a lava flow cooling model
• Answers to exercises
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
b
3
2 plinian
hawaiian
4 to 64 mm. This coarseness reflects the relatively
1
inefficient fragmentation of the liquid magma by
large gas bubbles, resulting from low liquid vis-
10–2 strombolian
cosities allowing relatively easy phase separation
(Chapter 4). Nevertheless, a range of particle sizes
0 10 20 30
is usually produced, a small proportion of which
Area1/2 (km)
can be sufficiently fine-grained to be lofted into
weakly buoyant thermal-like plumes. Figure 6.2╇╇ (a) Contrasting strombolian (lower, coarser, lighter
The strombolian eruptive process thus repre- colored) and hawaiian (upper, darker, finer-grained) products
of the eruption of Crater Hill volcano in the Auckland volcanic
sents the arrival, and subsequent bursting, of one
field, New Zealand.The better bedded nature of the strombolian
or more large (sizes approaching that of the con-
phase reflects its origin in numerous discrete short-lived
duit diameter), overpressured gas bubbles at the explosive events.The hawaiian deposits are dominated by
magma free surface. During periods of activity, microvesicular and highly fluidal lapilli whereas the strombolian
intervals between bursts are typically from tens clasts are more coarsely vesicular and ragged. Photo taken by
of minutes to multiple hours. However, eruptive thinning relationship requires, a priori, a con-
episodes can also involve periods of nearly con- stant release height of clasts, whereas the
tinuous bubble bursts, with intervals of only sec- heights of strombolian jets tend to vary greatly.
onds, producing persistent weak plumes of fine Consequently, strombolian and hawaiian erup-
ejecta and cascades of spatter. This represents a tions cannot be distinguished by average dis-
transitional type of activity which, from a classi- charge rate or ejecta dispersal; consideration of
fication point of view, merges into hawaiian-style the steadiness of discharge rate is required, with
lava fountaining. In terms of landforms, the low the former being impulsive, short-lived, discrete
dispersion and small volumes of coarse strom- events and the latter expressed as sustained
bolian ejecta determine that small, relatively fountaining, during which mass discharge rates
localized spatter- and scoria-dominated cones are are maintained for hours to days. So, although
formed around vents. However, in many cases, we concentrate here on classical discrete strom-
the style of activity varies between fountaining, bolian events involving individual large gas
transitional and pure strombolian, and signifi- bubbles, the two-phase flow behavior involved
cant cones (hundreds of meters high) can be built is closely linked to that of hawaiian-style erup-
by the greater magma effusion rates involved in tions. Transitional behavior (Parfitt and Wilson,
transitional and fountaining phases (Fig. 6.1(b)). 1995), and the conditions leading up to it are
A key issue is what parameter or parameters discussed in Section 4.8.3.
can be used to distinguish between these erup- The relatively high event frequency and
tive styles. Unlike transitions between many low-hazard nature of strombolian activity
other types of eruption, time averaged mass dis- has allowed intensive geophysical study of
charge rates cannot be used; for example, the 99 the processes involved. Many different data
eruptions or eruption episodes at Kilauea since types including seismic, imaging (both still
1955 show a range of time-averaged discharge and video, visible and thermal), infrasonic,
rates between 4 × 102 and 106 kg s–1, overlapping Doppler radar and sodar (sound detection and
completely with the rates displayed by strom- ranging), have been collected in order to ana-
bolian behavior (Houghton and Gonnermann, lyze ejecta velocities and masses, bubble over-
2008). Both strombolian and hawaiian eruptions pressures and sizes, eruption intermittencies,
produce cones composed of proximal pyroclast and to model depths of slug acceleration and
fall and, unlike Plinian falls, deposits do not fol- fragmentation. Consequently, when compared
low simple exponential thinning relationships with other types of less frequent and more
with distance from the vent (Fig. 6.2). This is per- hazardous volcanic activity, there is a signifi-
haps predictable, because a simple exponential cant and rapidly growing volume of data from
which models of strombolian activity can be
Caption For figure 6.2 continued╇╇ Colin Wilson. (b) Plot developed and tested. Furthermore, in contrast
of median diameter versus sorting coefficient for fall deposits to other scenarios in which magma rheological
from basaltic explosive eruptions modified after Walker and properties shift over many orders of magni-
Croasdale (1972) and Houghton et al. (2000) to include plinian tude (Chapter 4) and pyroclasts can exceed the
data from Carey et al. (2007). (c) Deposit thinning relationships sound speed and generate shocks (Chapter 7),
indicated by plots of tephra thickness versus the square root
strombolian eruptions are relatively straight-
of the area enclosed within a given isopach for selected basaltic
forward to model. For example, the essential
pyroclastic deposits. Strombolian deposits are represented by
open symbols (squares from Walker and Croasdale (1972));
subsurface fluid flow processes can be reason-
open triangles represent recent Arenal strombolian fall deposits ably represented using a liquid phase of con-
from Cole et al. (2005); closed triangles and bold line is the Etna stant viscosity. Nevertheless, in common with
2003 paroxysm.The hawaiian-style 1959 Kıˉlauea Iki eruption is most two-phase flows, a complete understand-
shown with closed squares and basaltic Plinian eruptions with ing remains elusive, and current quantitative
solid circles.Where only isomass data were available, they were models are generally limited to distinct regions
converted into thickness equivalents using field measurements of the overall process; the initial formation of
of bulk density for the deposits.
the large gas bubble, its ascent and burst, and
the subsequent dispersion of the pyroclasts. number densities of ~1010 to 1011 m–3 which, in a
These divisions are retained here and form the Newtonian liquid, have buoyant rise velocities,
ensuing structure of the chapter. The value of ub, that can be approximated by Stokes’ law
such models is that they allow our understand-
ing of the unobservable subsurface processes
ub =
( )
2 g ρ − ρ g rb 2
, (6.1)
that control eruptive style to be tested, using a 9η
wide variety of surface measurements. Not only
does this provide a starting point from which where g is the acceleration due to gravity, ρ
increasingly complex models of more hazard- and ρg are the liquid magma and gas densities,
ous eruptive phenomena can be derived but, respectively, rb is the bubble diameter, and η is
with the volumetric importance of basaltic vol- the magma viscosity. (All notation is given in
canism, it provides insight into the degassing Section 6.9). For considering the terminal rise
processes that continue to shape much of our velocities of bubbles in basalt magma, note
Earth and atmosphere. that Stokes’ law is derived for rigid spheres and
assumes creeping (laminar) flow within an infin-
ite incompressible Newtonian fluid. Thus, it is
only appropriate to use Eq. (6.1) for small (e.g.,
6.2╇ ╇ Slug-bubble formation radii of order 1 cm or less), widely separated
bubbles. At a pressure of 200 MPa and tempera-
As magma ascends, decompression drives the ture of ~1300 K, exsolved CO2 is a supercritical
process of gas exsolution and bubble expan- fluid1 with a density of ~100 kg m–3. So, for a
sion. Consequently, in order to understand magma viscosity of 100 Pa s, bubbles of radii
the transition from continuous subsurface between 10 µm and 1 mm have rise velocities
magma degassing to the sequence of discrete of 0.01 to 10 µm s–1, sufficiently slow that bub-
events observable at the surface, a fundamen- ble motion will be dominated by the magma
tal aspect of modeling strombolian activity is flow. Nevertheless, the bubbles will expand (and
to consider how individual, large bubbles can grow by diffusion) as they rise and depressurize
form from an evolving distribution of many and, when they reach radii of ~1 to 5 cm, ascent
small ones. In terms of two-phase flow, this velocities will have increased to ~1 to 10 cm s–1.
represents a transition from a “bubbly” to a These velocities are on the order of anticipated
“slug” flow regime (although neither need be magma velocities, so that the flow can be con-
in a steady state), where a “slug” is a large bub- sidered to be separated, with bubbles having the
ble of diameter similar to, and length larger chance to interact with each other or with the
than, the width of the conduit (Wallis, 1969; conduit walls.
see Chapter 4). Two models have been proposed for the bub-
At depth, the first volumetrically important ble interactions that underpin explosive basaltic
species to exsolve from the magma is CO2 and, eruptions. Using different conduit geometries,
even at depths of several kilometers, most CO2 the models differ in their consideration of
will be exsolved (Burton et al., 2007b; see Section where such interactions and the ensuing bub-
4.2.1). H2O is the principal magmatic volatile ble coalescence occur (Fig. 6.3). First, in the “rise
by weight but, being more soluble than CO2 it speed dependent” model (Wilson and Head,
exsolves at shallower depths (e.g., for Stromboli 1981), bubbles coalesce freely as they ascend a
magma at a depth of ~1.5 km, probably >€50% vertical conduit and eruptive style depends on,
of the magma water content is exsolved among other factors, the relative ascent rates of
(Burton et al., 2007b). The degassing process the bubbles and magma. In contrast, the “foam
begins with the nucleation of small bubbles, at collapse” model (Jaupart and Vergniolle, 1989;
1
╇A supercritical fluid is a fluid at a temperature and pressure above its critical point; for CO2 the critical point is at 304.1 K
and 7.38 MPa. Supercritical fluids tend to have physical properties in between those of a liquid and a gas.
Vergniolle and Jaupart, 1990) assumes bubbles of hawaiian fountains as the mixture of gas
collect and interact at major structural barri- and pyroclasts exits the vent. One of the main
ers such as magma chamber roofs and, under applications of the rise speed model has been to
this scenario, the effects of any magma motion address transitions between effusive, strombol-
are not considered. The merits and drawbacks ian, and hawaiian activity (Parfitt and Wilson,
of both models are discussed in detail by Parfitt 1995). For example, magma rise speed has a
(2004). strong influence on eruptive style but, for a rise
speed sufficiently low for strombolian activity
6.2.1â•… Rise speed dependent model to be maintained, this can change to hawai-
The rise speed dependent model (Fig. 6.3(a)) con- ian if the magma volatile content decreases,
siders the ascent rate of bubbles within magma reflecting the fact that the lower gas content
with respect to the ascent velocity of the magma no longer allows sufficient coalescence to gen-
as a whole, within a long, vertical conduit. erate the slug flow regime, but still permits
Larger bubbles ascend more rapidly (Eq. 6.1) decompression-induced fragmentation and the
and hence can encounter and absorb smaller formation of hawaiian fountains.
bubbles at an increasingly fast rate. If the ascent Although we address this model no further
velocity of the magma is low, then such bubbles here, one point is worth discussion. In the devel-
have time for sufficient coalescence to form gas opment of the rise speed model, a relatively high
slugs before the magma free surface is reached, gas volume fraction (0.75) was used to define the
generating strombolian activity. Thus, a depres- transition from bubbly to slug flow (Wilson and
surizing (ascending) bubbly flow, represents a Head, 1981). In contrast, engineering research
transitional flow regime that will develop into suggests that the bubble-to-slug transition gen-
slug flow if sufficient time is available and, in erally occurs as the gas volume fraction exceeds
long conduits, depressurization expansion will ~0.25 (Taitel et al., 1980). This has yet to be experi-
enhance this process. In contrast, relatively mentally verified in large diameter pipes (e.g.,
high magma rise speeds inhibit significant bub- > 100 mm) because, in the low-viscosity hydro-
ble coalescence prior to arrival at the surface, at carbons and aqueous solutions typically used in
which point, decompression-induced expansion laboratory simulations (η < 10–3 Pa s), turbulence
may ultimately lead to fragmentation of the disrupts large bubbles and hinders the develop-
homogeneous, bubbly flow and the formation ment of a slug flow regime (Omebere-Iyari et al.,
2007); experiments using high-viscosity liquids 0.69 was used for ε, which, for magma, implies
are required to study the transition at volcanic a foam thickness of < 10 cm for 1 mm bubbles,
scale. Nevertheless, large slugs can be observed increasing to tens of meters for micron-sized
transiently (before turbulence builds up) and, in bubbles (Jaupart and Vergniolle, 1989).
magmatic systems, despite even larger diameter Equating these two foam thicknesses, hm
conduits, excessive turbulence is unlikely due and hc, relates the input gas flux to a foam
to viscosities being up to six orders of magni- breakdown criterion. At fluxes less than the
tude greater. Consequently, there should be no critical flux, foam growth is steady and gas loss
reason why large gas bubbles will be unstable in is maintained by foam flow out of the conduit.
magma-filled conduits and slug flow may well Above the critical gas flux, the system moves
be possible at gas volume fractions much less into a cyclical regime in which the foam grows
than 0.75. to its maximum thickness then collapses into
a large gas pocket which rises rapidly into
6.2.2â•… Collapsing foam model the conduit, allowing the cycle to restart. The
In the collapsing foam model (Fig. 6.3(b)), bub- expression of this event at the free surface
ble interactions are forced by bubbles collect- depends on the duration of the collapse itself.
ing against a horizontal barrier, i.e., the roof of When the duration of the foam collapse is
a magma reservoir, hence coalescence can be short with respect to foam flow timescales, a
viewed as being directly driven by conduit geom- very large bubble can be produced and erup-
etry. Laboratory experiments in which small tions are deemed analogous to hawaiian-style
ascending bubbles formed a foam layer under a fountaining. However, when foam collapse is
flat roof (vented by only a small vertical upward relatively slow compared with foam flow, then
conduit) have shown that this system can exhibit numbers of small gas pockets are intermittently
either a steady or intermittent behavior (Jaupart produced and released, in a manner equivalent
and Vergniolle, 1989). As the foam grows, the to strombolian eruptions. In contrast to the
thickness, hm, of its deepest regions can be cal- rise speed model, in this model, increasing gas
culated by considering two gas fluxes: the input fluxes will convert strombolian-style activity
flux from bubbles from below and gas loss from to hawaiian as foam collapse becomes increas-
the foam as it flows into the open conduit ingly vigorous and continuous. Note that the
large horizontal roof required for foam collec-
1
ηf Q 4 tion is highly unlikely to exist in the relatively
h m = Ω 2 , (6.2)
ε ρg near-surface, thus initiation of these eruptions
will be controlled by CO2-dominated foams,
where ε is the gas volume fraction in the foam, ηf because significant H2O is not exsolved at mul-
is the foam viscosity, Q is the volume flux of gas, tiple-kilometer depths.
and Ω is a factor determined by the roof area
over which bubbles are collecting. However, 6.2.3â•… A combined model
a foam layer has a maximum potential thick- The rise speed and collapsing foam models
ness governed by the ability of surface tension illustrate well the effects of different bubble
to maintain the integrity of the foam against coalescence scenarios, but it is likely that nei-
increasing buoyancy forces. At the point of fail- ther of the two idealized conduit geometries,
ure, the critical foam thickness is given by invoked (a long, uniform vertical conduit or a
flat, horizontal magma chamber roof ) accur-
2σ
hc = , (6.3) ately represents a natural system. Furthermore,
ερ grb
the foam collapse model does not involve the
effects of bubble expansion and the rise speed
where σ is the surface tension. Experimental model does not include any “forced” coales-
results were best represented when a value of cence (such as is induced by a geometric trap).
10
–2 Eötvös numbers which, in the case of gas–liquid
0.2 0 flows where the gas density is sufficiently small
10 2
10 that it can be neglected, are
4
10
0.1 us
Fr = , (6.4)
gD
0.0
10 100 1000 gη4
Eötvös number, Eo Mo = , (6.5)
ρ σ3
Nf = = . (6.8)
Mo η rounding the slug (Batchelor, 1967)
1/ 3
In inclined pipes, the slug ascends along the 3ηrc u s
rs = rc − . (6.10)
upper region of the pipe wall, allowing a thicker 2ρ g
return flow of liquid under the slug body.
Consequently, ascent velocities in inclined For a slug with initial length ls0, within a ver-
pipes can exceed those in vertical pipes by up tical cylindrical conduit, conservation of liquid
3.10
1.82
1.09
0.73
(mg)
0.37
surface
( ) ( )
h0 rc2 + l s 0 rc2 − rs 2 = ( h + u s t ) rc2 + l s rc2 − rs 2 , (6.11) 2.0
acceleration of the slug nose towards the sur- Shear force magnitude (×10 N m )
5 –1
face. For example, in a 1.5 m radius conduit 0 0.1 0.4 0.9 1.6 2.5
and magma of density 2600 kg m–3 and vis-
cosity 500 Pa s, the average slug nose velocity
20
ascending through the last 10 m of magma is a
Force (×106 N)
force 1
the dynamics of the fluid surrounding the slug
0
are not included, forces and pressures exerted
shear –1
on the conduit cannot be calculated accurately. force
Thus, in order to interpret seismic data (Chapter
Force (×106 N) 1 c
16), more sophisticated fluid dynamic models net force
are required to fully describe the liquid motion.
0
Some computational fluid dynamic (CFD) mod-
els of slug flow have been developed through 0 10 20 30 40
engineering research (Mao and Dukler, 1990; Time (s)
Clarke and Issa, 1997; Taha and Cui, 2006); how- Figure 6.6╇╇The vertical force components exerted on
ever, they only describe the local flow around a conduit during the ascent and burst of a gas slug, as
non-expanding gas slugs and are thus restricted determined by 3D fluid dynamic modeling. The scenario
in their direct application to volcanological simulated has a gas mass of 70 kg, a conduit radius of 1.5 m
scenarios. and a magma density and viscosity of 2600 kg m–3 and 500
In order to simulate flow in a complete con- Pa s, respectively. In (a) the force per vertical meter of the
duit, classic finite difference computational conduit exerted by viscous shear on the conduit wall is shown
over the duration of slug ascent. The position of the magma
fluid dynamics (CFD) and Lattice Boltzman
surface and the slug nose and base are shown by the black
numerical models are currently being explored
curves, with the former two truncated ~1 s before slug burst
(O’Brien and Bean, 2008; D’Auria and Martini, for clarity. In (b) the total vertical forces over the domain due
2009), with the results being linked to the gen- to pressure and shear forces are shown, and are summed in
eration of seismic signals. A similar approach (c). The slug burst is accompanied by a sharp upward force on
has also been taken using the commercial com- the domain. Redrawn from James et al. (2008) with permission
putational fluid dynamics code Flow-3D®, and from the Geological Society of London.
comparing the results with laboratory simula-
tions (James et al., 2008). Flow-3D® solves the
Navier–Stokes equations over a fixed grid and be adequately modeled by either PsVs or PsVsγ =
specializes in free-surface flows. For the ascent constant, representing isothermal or adiabatic
of a gas slug, the large density and viscosity expansion, respectively.
contrasts between the gas and liquid phases A detailed discussion of the model is beyond
allow the internal motion of the gas phase to the scope of this chapter but the results give
be neglected; the position of the gas–liquid insight into forces exerted on the conduit dur-
interface is controlled by the liquid and gas ing slug expansion, in particular, those pro-
pressures, the liquid dynamics, and surface ten- duced by viscous shear. Figure 6.6 illustrates
sion (if significant). The slug gas pressure can the shear forces exerted on a vertical conduit,
with the position of the slug indicated by the quickly in viscous magma and may not be excited
ascending region of downward shear exerted at all if the slug expansion occurs too rapidly to
by the falling liquid film around the slug. This enable a resonance to become established. Note
downward shear applies a region of vertical ten- that these pressure changes are much lower in
sion in the conduit walls immediately above the amplitude than the overall changes in magma-
slug nose and a compressional region behind static pressure due to the passage of a slug but,
the slug base. As the slug begins to decompress with frequency ranges (of around a few Hertz)
rapidly near the surface, the upward acceler- within the response bands of seismic and acous-
ation of the liquid above the slug exerts a strong tic techniques, they lend themselves to being
upward force on the conduit. Extension of this detected.
type of model to more complex conduit geom-
etries may allow some seismic signals (e.g., very
long period events) to be interpreted in terms of 6.4╇ ╇ Burst of a gas slug
fluid dynamic processes (see Chapter 15). This
raises the possibility that even in relatively low When the bubble nose reaches the magma sur-
viscosity magmas, the effect of viscous drag may face, a thinning magma membrane is produced
be detectable, just as has been hypothesized for which blisters upwards before bursting to prod-
much more viscous magmas, for example at uce the observable explosion. Some infrasonic
Asama, Soufrière Hills, and Mount St. Helens waveforms have been interpreted as oscillations
volcanoes. of the membrane prior to bursting (Vergniolle
and Brandeis, 1996; Vergniolle et al., 1996).
6.3.4â•… Near-surface oscillations However, within conduits, viscous coupling to
As slugs approach the surface, seismic and the conduit wall swiftly removes liquid from
acoustic data record a variety of signals (see above rapidly expanding slugs, hence viable
Chapters 15 and 16), some of which have been membranes will be extremely short lived tran-
interpreted in terms of pressure oscillations sients, and oscillations are very unlikely. For
within the gas slugs (Vergniolle and Brandeis, unconfined gas bubbles bursting at the open
1996; Vergniolle et al., 1996). Different oscilla- surface of a lava lake, this wall interaction does
tory modes have been proposed: longitudinal not occur and membrane oscillations could be
compression and dilation of the slug, with a considered. Nevertheless, no such oscillations
frequency given by have been reported from video and radar data
of large bubbles bursting in lava lakes (Dibble
1 γ Ps et€ al., 2008; Gerst et al., 2008) which, instead,
fL = , (6.15)
2π ρl s h indicates that bubble overpressures alone drive
sufficient fluid accelerations for rapid disruption
and kinematic waves forming at the slug nose, (fragmentation).
with a frequency The criteria for predicting when the mem-
brane will fail are difficult to quantify precisely,
1
0.96 N 3 ρ g 2
3
but the burst process is important for interpret-
f NR0 = , (6.16)
8π 3η ing data such as ejecta velocity and infrasonic
transients. A plausible and straightforward
where N is the mode (N = 1 being the funda- first-order burst criterion is to define a minimum
mental mode), with odd modes being favored. thickness for the membrane before it ruptures
In laboratory experiments involving low rates and, with some of the ejecta being composed of
of slug expansion, pressure changes observed parts of this membrane, it is reasonable to relate
to increase in frequency (as h–½) as the slug the minimum membrane thickness to the size
approached the surface have been attributed to of the observed ejecta. In the case of Stromboli,
longitudinal slug oscillations (James et al., 2004). where characteristic particle dia�meters are
However, such oscillations may be damped very often 1 to 4 cm, a membrane thickness of order
dike segments, with dips ranging between 40° role of post-burst drainback and rainfall (which
and ~70° (Chouet et al., 2008). has been observed to affect eruption velocities
The wealth of such geophysical data from and pyroclast sizes as determined by Doppler
Stromboli illustrates the complexity of the sys- radar) are also not understood. Even fundamen-
tem with respect to the current first-order mod- tal questions such as where does the change
els of strombolian activity. Many key questions from a dike-like morphology to a circular con-
remain and are illustrated by inconsistencies duit occur, and what process is responsible for
between results derived from different tech- switching activity styles between vents, remain
niques. For example, modeling oscillatory acous- largely unaddressed.
tic data has suggested low magma viscosities of
~300 Pa s (Vergniolle et al., 1996). Although this
is not inconsistent with values successfully used
in fluid dynamic models of slug ascent, petrolog-
6.6╇ ╇Example: Mount Erebus,
ical modeling suggests that some magma could Antarctica
have a viscosity of up to 1.4 × 104 Pa s (Metrich
et al., 2001). A currently unexplored area is the A contrasting example is illustrated by activity
extent to which a non-Newtonian and spatially at Mount Erebus, where strombolian activity is
varying magma rheology would affect our cur- represented by bursts of large, observable bub-
rent understanding of slug ascent and burst bles within a small lava lake. Mount Erebus is a
processes. Further complexities such as the large stratovolcano located on Ross Island, at the
western end of the West Antarctic rift system, for the subsurface plumbing at Erebus involve
with a summit at 3794 m. The summit cone hosts a relatively large, high-level storage region, pos-
the Main Crater, within which the Inner Crater sibly of the diameter of the crater itself, such that
is approximately 100 m deep, with a radius of bubble coalescence would take place only in the
~80€m. Over the last 30 years, the Inner Crater has upper few tens of meters (Aster et al., 2003). The
contained an anorthoclase phonolite lava lake of lack of significant upwelling of the lake immedi-
between 5 and 15 m in radius that has been the ately prior to slug burst supports this interpret-
site of variable strombolian activity. Events are ation by suggesting there is only a brief period of
generally simple rather than complex, and gen- slug ascent. Upwelling may be also suppressed by
erally occur at a rate of several per day, with the a strongly flared upper region of the conduit.
occasional swarm of up to 900 per day. Erebus is During bursts, the surface of the lava lake
now permanently geophysically monitored, with is observed to dome hemispherically upwards
seismic, infrasonic, and video data being con- before being disrupted and showering the cra-
tinuously collected unless instruments succumb ter area with pyroclasts. Radar data collected
to the harsh environmental conditions. in 2005–2006 suggest that pressures at burst
Early work described upwelling lava in the could be ~3 × 105 to 8 × 105 Pa (Gerst et al.,
lake accompanied by seismic tremor preceding 2007a). Radar (Gerst et al., 2008) and infrasonic
strombolian events by ~20 to 45 s. Erupting bub- (Jones et al., 2008) data indicate that the pos-
bles would have radii of up to 5 m, after which ition (within the lava lake) and initial direction
lake levels would drop by about 5 m before of the burst are random, strongly suggesting
recovering over the following 15 minutes. This that, close to the surface, the conduit is likely
type of activity is indicative of slugs expanding to be vertical and that the burst process itself
as they ascend and, assuming a slug velocity is controlled by instabilities or heterogeneities
between 1 and 2 m s–1, suggests a minimum slug within the expanding magma membrane. With
source depth of between 20 and 90 m. the membrane composed of magma that has
More recently, VLP seismic events have been been rapidly and recently cooling in the atmos-
recorded and, although depths could be exagger- phere, this is highly likely.
ated by the effects of tilt, particle motions suggest Erebus provides a good case for multipa-
VLP source depths of between 100 and 800 m rameter (video, radar, infrasonic, seismic)
(Rowe et al., 1998, 2000). Gas geochemistry data are measurements of bubble bursts. However, the
also consistent with gas being sourced at depths subsurface geometries are poorly constrained
of up to 2 km (Oppenheimer and Kyle, 2008). and key questions remain, such as how to rec-
However, the VLP events precede short-period sig- oncile the contrasting evidence for the depth of
nals (which accompany the surface explosion) by slug coalescence.
up to only 5 s, suggesting very short slug ascents,
of order 10 m. This is in agreement with video
data that indicate only the very smallest of distur- 6.7╇ ╇ Future directions
bances to the lake level prior to each event and
no significant period of decompression-related Our understanding of strombolian activity has
gas expansion. The relative timings of the seis- been advancing dramatically, most recently due
micity and the polarity of the VLP signals vary, to quantitative measurement of the gas species
suggesting multiple slug ascent paths within a from spectroscopic data and improved imaging
well-connected plumbing system. Furthermore, it of conduit segments by inverted VLP seismic
is likely that the location of the VLP source reflects data. Such findings have been augmented by
the region of greatest coupling between the vary- additional key interpretations drawn from an
ing fluid pressures and the surrounding rock increasingly diverse array of other investiga-
rather than the locus of the initial fluid-dynamics tions, including laboratory experiments, com-
driven pressure perturbation (Aster et al., 2003; putational fluid dynamic modeling, thermal
James et al., 2006). Consequently, current models imaging, and infrasonic data collection.
us slug ascent velocity, defined at the slug Carey, R. J., Houghton, B. F., Sable, J. E. and
base (m s–1) Wilson, C. J. N. (2007). Contrasting grain size and
Vs volume of a gas slug (m3) componentry in complex proximal deposits of the
γ ratio of specific heats 1886 Tarawera basaltic plinian eruption. Bulletin of
ε foam gas volume fraction Volcanology, 69, 903–926.
Chouet, B., Hamisevicz, N. and McGetchin, T. R. (1974).
η magma viscosity (Pa s)
Photoballistics of volcanic jet activity at Stromboli,
η f foam viscosity (Pa s)
Italy. Journal of Geophysical Research, 79, 4961–4976.
ρ liquid magma density (kg m–3)
Chouet, B., Dawson, P., Ohminato, T. et al.
ρg gas density (kg m–3)
(2003). Source mechanisms of explosions at
ρp pyroclast density (kg m–3) Stromboli Volcano, Italy, determined from
σ magma surface tension (N m–1) moment-tensor inversions of very-long-period
Ω area factor for bubble collection under a data. Journal of Geophysical Research, 108, 2019,
horizontal roof doi:10.1029/2002JB001919.
Eo Eötvös number Chouet, B., Dawson, P. and Martini, M. (2008).
Fr Froude number Shallow-conduit dynamics at Stromboli Volcano,
Frθ Froude number at an inclination angle θ Italy, imaged from waveform inversions. In Fluid
Mo Morton number Motion in Volcanic Conduits: A Source of Seismic and
Acoustic Signals, ed. S. J. Lane and J. S. Gilbert.
Geological Society, London, Special Publication
307, pp. 57–84.
References Clarke, A. and Issa, R. I. (1997). A numerical model of
slug flow in vertical tubes. Computers and Fluids, 26,
Allard, P. (2010). A CO2-rich gas trigger of explosive 395–415.
paroxysms at Stromboli basaltic volcano, Italy. Cole, P. D., Fernandez, E., Duarte, D. and
Journal of Volcanology and Geothermal Research, 189, Duncan,€A.€M. (2005). Explosive activity and
363–374, doi:10.1016/j.jvolgeores.2009.11.018. generation mechanisms of pyroclastic flows at
Aster, R., Mah, S., Kyle, P. et al. (2003). Very long Arenal volcano, Costa Rica between 1987 and
period oscillations of Mount Erebus Volcano. 2001. Bulletin of Volcanology, 67, 695–716.
Journal of Geophysical Research, 108, 2522, D’Auria, L. and Martini, M. (2009). Slug flow:
doi:10.1029/2002JB002101. Modeling in a conduit and associated elastic
Barberi, F., Rosi, M. and Sodi, A. (1993). Volcanic radiation. In Encyclopedia of Complexity and Systems
hazard assessment at Stromboli based on review of Science, ed. R. A. Meyers, pp. 8153–8168.
historical data. Acta Vulcanologica, 3, 173–187. Dibble, R. R., Kyle, P. R. and Rowe, C. A. (2008).
Batchelor, G. K. (1967). An Introduction to Fluid Video and seismic observations of Strombolian
Dynamics. Cambridge: Cambridge University Press. eruptions at Erebus volcano, Antarctica. Journal of
Blackburn, E. A., Wilson, L. and Sparks, R. S. J. (1976). Volcanology and Geothermal Research, 177, 619–634,
Mechanics and dynamics of Strombolian activity. doi:10.1016/j.jvolgeores.2008.07.020.
Journal of the Geological Society of London, 132, 429–440. Dubosclard, G., Donnadieu, F., Allard, P. et al. (2004).
Burton, M., Allard, P., Mure, F. and La Spina, A. Doppler radar sounding of volcanic eruption
(2007a). Magmatic gas composition reveals the dynamics at Mount Etna. Bulletin of Volcanology, 66,
source depth of slug-driven Strombolian explosive 443–456.
activity. Science, 317, 227–230. Gerst, A., Hort, M., Johnson, J. B. and Kyle, P. R.
Burton, M. R., Mader, H. M. and Polacci, M. (2007b). (2007a). The dynamics of a Strombolian bubble
The role of gas percolation in quiescent degassing burst derived from Doppler radar. Eos, Transactions
of persistently active basaltic volcanoes. Earth and of the American Geophysical Union, 88, Fall Meeting
Planetary Science Letters, 264, 46–60, doi:10.1016/j. Supplement, Abstract V24B-06.
epsl.2007.08.028. Gerst, A., Hort, M., Johnson, J. B. and Kyle, P. R.
Calvari, S. and Pinkerton, H. (2004). Birth, growth (2007b). The first second of a Strombolian eruption:
and morphologic evolution of the ‘Laghetto’ Doppler radar and infrasound observations at
cinder cone during 2001 Etna eruption. Journal of Erebus volcano, Antarctica. Geophysical Research
Volcanology and Geothermal Research, 132, 225–239. Abstracts, 9, 07280.
Gerst, A., Hort, M., Kyle, P. R. and Vöge, M. (2008). 4D Menand, T. and Phillips, J. C. (2007). Gas segregation
velocity of Strombolian eruptions and man-made in dykes and sills. Journal of Volcanology and
explosions derived from multiple Doppler Geothermal Research, 159, 393–408.
radar instruments. Journal of Volcanology and Metrich, N., Bertagnini, A., Landi, P. and Rosi, M.
Geothermal Research, 177, 648–660, doi:10.1016/j. (2001). Crystallization driven by decompression
jvolgeores.2008.05.022. and water loss at Stromboli volcano (Aeolian
Houghton, B. F. and Gonnermann, H. M. (2008). Islands, Italy). Journal of Petrology, 42, 1471–1490.
Basaltic explosive volcanism: constraints from Metrich, N., Bertagnini, A., Landi, P., Rosi, M. and
deposits and models. Chemie der Erde, 68, 117–140. Belhadj, O. (2005). Triggering mechanism at
Houghton, B. F., Wilson, C. J. N. and Smith, I.€E.€M. the origin of paroxysms at Stromboli (Aeolian
(2000). Shallow-seated controls on styles of Archipelago, Italy): The 5 April 2003 eruption.
explosive basaltic volcanism: a case study from Geophysical Research Letters, 32, L10305.
New Zealand. Journal of Volcanology and Geothermal Metrich, N., Bertagnini, A. and Di Muro, A. (2010).
Research, 91, 97–120. Conditions of magma storage, degassing and
James, M. R., Lane, S. J., Chouet, B. and Gilbert, J. S. ascent at Stromboli: New insights into the
(2004). Pressure changes associated with the ascent volcano plumbing system with inferences on the
and bursting of gas slugs in liquid-filled vertical eruptive dynamics. Journal of Petrology, 51, 603–626,
and inclined conduits. Journal of Volcanology and doi:10.1093/petrology/egp083.
Geothermal Research, 129, 61–82. O’Brien, G. S. and Bean, C. J. (2008). Seismicity
James, M. R., Lane, S. J. and Chouet, B. A. (2006). Gas on volcanoes generated by gas slug ascent.
slug ascent through changes in conduit diameter: Geophysical Research Letters, 35, L16308,
Laboratory insights into a volcano-seismic doi:10.1029/2008gl035001.
source process in low-viscosity magmas. Omebere-Iyari, N. K., Azzopardi, B. J. and Ladam, Y.
Journal of Geophysical Research, 111, B05201, (2007). Two-phase flow patterns in large diameter
doi:10.1029/2005JB003718. vertical pipes at high pressures. AIChE Journal, 53,
James, M. R., Lane, S. J. and Corder, S. B. (2008). 2493–2504, doi: 10.1002/aic.11288.
Modelling the rapid near-surface expansion of Oppenheimer, C. and Kyle, P. R. (2008). Probing the
gas slugs in low viscosity magmas. In Fluid Motion magma plumbing of Erebus volcano, Antarctica,
in Volcanic Conduits: A Source of Seismic and Acoustic by open-path FTIR spectroscopy of gas emissions.
Signals, ed. S. J. Lane and J. S. Gilbert : Geological Journal of Volcanology and Geothermal Research, 177,
Society, London, Special Publication 307, 743–754, doi:10.1016/j.jvolgeores.2007.08.022.
pp.€147–167. Parfitt, E. A. (2004). A discussion of the mechanisms
Jaupart, C. and Vergniolle, S. (1989). The generation of explosive basaltic eruptions. Journal of Volcanology
and collapse of a foam layer at the roof of a and Geothermal Research, 134, 77–107.
basaltic magma chamber. Journal of Fluid Mechanics, Parfitt, E. A. and Wilson, L. (1995). Explosive
203, 347–380. volcanic eruptions IX. The transition between
Jones, K. R., Johnson, J. B., Aster, R., Kyle, P. R. and hawaiian-style lava fountaining and strombolian
McIntosh, W. C. (2008). Infrasonic tracking of explosive activity. Geophysical Journal International,
large bubble bursts and ash venting at Erebus 121, 226–232.
Volcano, Antarctica. Journal of Volcanology and Patrick, M. R., Harris, A. J. L., Ripepe, M. et al.
Geothermal Research, 177, 661–672, doi:10.1016/j. (2007). Strombolian explosive styles and source
jvolgeores.2008.02.001. conditions: insights from thermal (FLIR) video.
Lautze, N. C. and Houghton, B. F. (2005). Physical Bulletin of Volcanology, 69, 769–784, doi:10.1007/
mingling of magma and complex eruption s00445–006–0107–0.
dynamics in the shallow conduit at Stromboli Polacci, M., Baker, D. R., Bai, L. P. and Mancini, L.
volcano, Italy. Geology, 33, 425–428. (2008). Large vesicles record pathways of degassing
Mao, Z. S. and Dukler, A. E. (1990). The motion at basaltic volcanoes. Bulletin of Volcanology, 70,
of Taylor bubbles in vertical tubes 1. A 1023–1029, doi:10.1007/s00445–007–0184–8.
numerical-simulation for the shape and rise Ripepe, M., Rossi, M. and Saccorotti, G. (1993).
velocity of Taylor Bubbles in stagnant and Image-processing of explosive activity at
flowing liquid. Journal of Computational Physics, 91, Stromboli. Journal of Volcanology and Geothermal
132–160. Research, 54, 335–351.
Rosi, M., Bertagnini, A., Harris, A. J. L. et al. (2006). Weill, A., Brandeis, G., Vergniolle, S. et al. (1992).
A case history of paroxysmal explosion at Acoustic sounder measurements of the vertical
Stromboli: Timing and dynamics of the April 5, velocity of volcanic jets at Stromboli Volcano.
2003 event. Earth and Planetary Science Letters, 243, Geophysical Research Letters, 19, 2357–2360.
594–606. White, E. R. and Beardmore, R. H. (1962). The
Rowe, C. A., Aster, R. C., Kyle, P. R., Schlue, J. W. velocity of rise of single cylindrical air bubbles
and Dibble, R. R. (1998). Broadband recording through liquids contained in vertical tubes.
of Strombolian explosions and associated Chemical Engineering Science, 17, 351–361.
very-long-period seismic signals on Mount Erebus Wilson, L. (1980). Relationships between pressure,
volcano, Ross Island, Antarctica. Geophysical Research volatile content and ejecta velocity in three types
Letters, 25, 2297–2300. of volcanic explosion. Journal of Volcanology and
Rowe, C. A., Aster, R. C., Kyle, P. R., Dibble, R. R. Geothermal Research, 8, 297–313.
and Schlue, J. W. (2000). Seismic and acoustic Wilson, L. and Head, J. W. (1981). Ascent and
observations at Mount Erebus Volcano, Ross Island, eruption of basaltic magma on the Earth and
Antarctica, 1994–1998. Journal of Volcanology and Moon. Journal of Geophysical Research, 86, 2971–3001.
Geothermal Research, 101, 105–128. Zukoski, E. E. (1966). Influence of viscosity, surface
Seyfried, R. and Freundt, A. (2000). Experiments on tension and inclination angle on motion of long
conduit flow and eruption behavior of basaltic bubbles in closed tubes. Journal of Fluid Mechanics,
volcanic eruptions. Journal of Geophysical Research, 25, 821–837.
105, 23727–23740.
Steinberg, G. S. and Babenko, J. J. (1978).
Experimental velocity and density determination Exercises
of volcanic gases during eruption. Journal of
Volcanology and Geothermal Research, 3, 89–98.
6.1 Consider the ascent of small bubbles of H2O vapor
Taha, T. and Cui, Z. F. (2006). CFD modelling of slug
within a basaltic magma conduit.
flow in vertical tubes. Chemical Engineering Science,
(a) Estimate the rise velocity through the fluid, ub,
61, 676.
of bubbles with diameters of 100 μm, 1 mm, 1
Taitel, Y., Barnea, D. and Dukler, A. E. (1980).
cm, and 10 cm, stating your assumptions.
Modelling flow pattern transitions for steady
(b) Stokes’ law is valid only for laminar flow
upward gas-liquid flow in vertical tubes. AIChE
described by Reynolds numbers, Re, less than
Journal, 26, 345–354, doi:10.1002/aic.690260304
approximately 0.3, where Re = ρubL/η, ρ and
Vergniolle, S. and Jaupart, C. (1990). Dynamics of
η are the fluid density and viscosity respect-
degassing at Kilauea Volcano, Hawaii. Journal of
ively, and L is a characteristic length (e.g.,
Geophysical Research, 95, 2793–2809.
bubble diameter). In this case, comment on
Vergniolle, S. and Brandeis, G. (1996). Strombolian
the validity of your answers to (a).
explosions 1. A large bubble breaking at the
(c) Would anticipated rise velocities be less or
surface of a lava column as a source of sound.
greater than calculated, and why?
Journal of Geophysical Research, 101, 20433–20447.
Vergniolle, S., Brandeis, G. and Mareschal,€J. C. 6.2 As gas ascends within conduits, it expands due to
(1996). Strombolian explosions 2. Eruption the decreasing pressure.
dynamics determined from acoustic (a) For a bubble of volume 1 m3 at a depth of 1000
measurements. Journal of Geophysical Research, 101, m in a basaltic conduit, estimate and plot the
20449–20466. bubble radius against depth, assuming that
Viana, F., Pardo, R., Yanez, R., Trallero, J. L. and the bubble could ascend to the surface in
Joseph, D. D. (2003). Universal correlation for the spherical form. What is the bubble volume
rise velocity of long gas bubbles in round pipes. once erupted?
Journal of Fluid Mechanics, 494, 379–398. (b) If the bubble ascends a conduit as a slug of
Walker, G. P. L. and Croasdale, R. (1972). near-cylindrical shape and constant radius 1.2
Characteristics of some basaltic pyroclastics. m, what length would the slug achieve at the
Bulletin Volcanologique, 35, 303–317. surface?
Wallis, G. B. (1969). One-Dimensional Two-Phase Flow. (c) Discuss whether, if this slug were “real,” it
New York: McGraw-Hill. would reach this length or not; what is the
evidence and what assumptions and omis- (c) Considering the shape of the pyroclast tracks,
sions from the calculations result in the what is the primary control on their trajector-
differences? ies and what other factors must also be influ-
(d) Conditions appropriate to basaltic activity encing their motion?
on Io or the early Moon would include very (d) Given your answers to (c), explain why
low atmospheric pressures and a much lower Strombolian deposits appear significantly less
gravity compared to that on Earth. Consider dispersed than Plinian deposits (Fig. 6.2(c)).
these extraterrestrial environments by repeat-
ing your calculations for parts (a–c) using, for 6.4 Data on the deposits from two well-documented
simplicity, a surface atmospheric pressure historical eruptions are shown with filled sym-
of 1 Pa and a gravitational acceleration of bols in Figure. 6.2(c), from the 1959 hawaiian
1€ m s-2. Discuss the differences between the fountaining eruption of Kilauea (filled squares)
results and those for terrestrial conditions, and the 2003 paroxysm at Stromboli (filled
commenting on the implications for eruptive triangles).
style. (a) From the graph data alone, contrast these two
eruptions in terms of their intensity (eruption
6.3 Figure 6.1(a) shows a long-exposure photograph rate) and volume.
of a strombolian eruption. (b) What do your interpretations imply about
(a) As the gas slug arrived at the surface, what the relative duration of these eruptions?
pressure conditions must have existed inside (c) How then can you explain the thickness
the slug in order to generate and eject the values given for the Strombolian deposits
pyroclasts? with open squares, which overlap with the
(b) With respect to the shape and position of the data for the hawaiian fountaining eruption?
slug immediately prior to burst, where is the
material that forms the pyroclasts sourced Answers to exercises available at www.�cambridge.
from? org/fagents
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
is disrupted into a gas–pyroclast mixture, pro- although commonly plumes rise < 10 km (e.g.,
pelled upward, and ejected from the vent at Rose et€ al., 1978; Glaze et€ al., 1989; Stix et€ al.,
velocities up to 400 m s−1 (Self et€al., 1979). The 1997; Druitt et€al., 2002; Wright et€al., 2006). For
resulting eruption may produce buoyant col- comparison, cruising altitude for commercial
umns, pyroclastic density currents, or both. air traffic is generally 9−12 km. Vulcanian col-
Typically, only a portion of the magma in the umn heights typically exceed heights associated
conduit is fragmented and evacuated (e.g., Druitt with strombolian eruptions (Chapter 6) and are
et€al., 2002), such that vulcanian eruptions char- less than those associated with plinian and sub-
acteristically last only seconds to minutes. They plinian eruptions (Chapters 8 and 9).
may occur as single events or as a sequence of Vulcanian eruptions have a wide range of
explosions spaced sufficiently far apart in time dispersal areas and degrees of fragmentation,
to produce distinguishable, discrete, unsteady making them difficult to classify from deposit
jets, and plumes. characteristics alone (Fig. 7.2; Walker, 1973;
The short duration and unsteady vent con- Cas and Wright, 1987). Vulcanian fall deposits
ditions of vulcanian eruptions make them dis- are typically thin, may consist of clasts ranging
tinct from quasi-steady plinian or subplinian from fine ash to coarse lapilli, and contain vari-
eruptions, for which it is generally assumed able proportions of lithics (non-juvenile) and
that bubbly magma rises to meet a fragmenta- juvenile material depending on the specifics of
tion wave and thus steadily feeds vent flux over the eruption (Self et€ al., 1979; Cas and Wright,
long periods (hours to days). Vulcanian erup- 1987; Morrissey and Mastin, 2000). Juvenile
tions typically erupt < 0.1 km3 DRE (dense rock clast densities vary even within a single erup-
equivalent), corresponding to a mass of ~1011€kg tion, ranging from dense to highly vesicular
(Morrissey and Mastin, 2000), and produce (e.g., Formenti and Druitt, 2003; Clarke et€ al.,
columns < 20 km high (Nairn and Self, 1978), 2007). Furthermore, unlike strombolian events,
1997; Sparks, 1997; Moore and Carmichael, retard flow, must be balanced by the pressure
1998; Hammer et€al., 1999; Cashman and Blundy, gradient as follows
2000). Crystallization and degassing increase
dP 32η u
magma viscosity (Lejeune and Richet 1995; Hess = −ρ g − , (7.1)
dz D2
and Dingwell, 1996; Sparks, 1997; Stix et€ al.,
1997; Melnik and Sparks, 1999) and, as a result where P is the magma pressure, z is the vertical
of concentrating volatiles in the remaining melt, coordinate increasing upward, ρ and η are the
force further degassing. Bubble connections are density and viscosity of the ascending magma,
thought to develop at some threshold vesicu- g is gravitational acceleration, D is the dia�meter
larity (Eichelberger et€ al., 1986; Takeuchi et€ al., of the conduit, and u is the ascent velocity of
2005). However, simultaneous crystallization the magma (Melnik and Sparks, 1999). All
and degassing tends to concentrate vesicles into para�meters except D are allowed to vary with
the interstices between crystals (Fig.€7.3(a)), and z. The first term on the right-hand side repre-
may enhance permeable connections among sents gravitational forces while the second term
bubbles and lead to open-system degassing at represents viscous forces according to assump-
lower-than-expected porosities. The consequent tions associated with Poiseuille flow (steady, vis-
gas loss may cause vesicle collapse and magma cous flow through a pipe, Fox et€al., 2006). The
densification, resulting in a dense, coherent magma viscosity η is a function of composition,
plug which seals the conduit and may force temperature, dissolved water content, crystal
underlying magma and gas to stagnate (Fig.€7.3; content, and bubble content (Chapters 4 and 5).
Hammer et€ al., 1999; Taddeucci et€ al., 2004; Because most of the increase in magma
Cashman and McConnell, 2005; D’Oriano et€al., viscosity occurs near the surface (and because
2005; Diller et€al., 2006; Clarke et€al., 2007). the increase in the viscous term outweighs
Many of these same processes cause pres- the decrease in the gravitational term in Eq.
sure to increase in the shallow conduit. Two (7.1)), most of the pressure drop must occur
mechanisms are particularly relevant: magma in the upper conduit precisely where viscosity
pressure may increase due to rheological stiff- increases most rapidly (Sparks, 1997). In other
ening of the ascending magma mixture (Sparks words, the pressure gradient dP/dz between the
1997; Melnik and Sparks 1999) and volatile pres- chamber and the upper portions of the conduit
sure may increase because bubble growth is is gentle because the viscosity gradient is gen-
restricted by high-viscosity magma (Fig. 7.3(c)); tle in that region, whereas the pressure gradient
Sparks, 1978; Tait et€al., 1989; Stix et€al., 1997). in the top few hundred meters of the conduit
Pressure in the shallow conduit increases until is steep because the viscosity gradient is steep
the plug is dislodged at pressures as high as 10 in that region. This complex pressure gradient
MPa, based on the strength of typical magmas. allows high pressures to extend over a large por-
Disruption results in fragmentation of both the tion of the conduit (Fig. 7.3(c)).
plug and underlying magma (Fig. 7.1(b)). Bubble overpressure develops primarily due
Magma overpressure develops in the shal- to viscosity-limited growth of bubbles for magma
low conduit because rheological stiffening con- viscosities > 107 Pa s (Chapter 4; McBirney and
centrates most of the pressure gradient in the Murase, 1971; Sparks, 1978; Proussevitch and
upper portions of the conduit (Sparks, 1997; Sahagian, 1996; Navon and Lyakhovsky, 1998).
Melnik and Sparks, 1999). The total pressure The gas phase expands in response to decreas-
drop from magma chamber to surface drives ing ambient magma pressure; however, bubble
flow upward from the chamber into the upper growth is resisted by surrounding high-viscosity
conduit between eruptions. For the one-dimen- melt, resulting in a trapped high-pressure gas
sional steady-state, laminar flow case, and for phase. In general terms, bubble overpressure ΔP,
slow ascent rates in a conduit of constant dia� defined as the difference between the pressure
meter, gravitational and viscous forces, which in the volatiles contained within the bubbles
magma or surrounding hot rock and, in turning 1981; Turcotte et€al., 1990; Woods, 1995; Druitt
to steam, increases pressure beneath the plug et€al., 2002; Clarke et€al., 2002a,b). Therefore the
and dislodges it (Schmincke, 1977; Fagents and fragmentation front meets magma with varying
Wilson, 1993). This process is distinct from that degrees of crystallinity, vesicularity, and viscos-
of phreatomagmatic fragmentation. In the case ity, partly explaining the wide range of eruptive
of vulcanian eruptions, groundwater vaporiza- products found in vulcanian deposits.
tion simply serves to dislodge the plug, which Key questions about the specifics of frag-
forces the underlying magma to fragment via a mentation in vulcanian eruptions remain
decompression-induced fragmentation wave. unanswered. What is the precise mechanism of
fragmentation at the front? How can the process
be quantified? What causes the fragmentation
front to stop? Many models of fragmentation
7.3╇ ╇Decompression and have been put forth (Chapter 4) that may help
fragmentation to answer these questions. The three most
common and important are discussed here: (1)
Upon plug disruption, a decompression wave strain-rate threshold (Dingwell, 1996; Papale,
travels at the local sound speed into the con- 1999); (2) tensile stress threshold (Zhang, 1999;
duit (Kieffer, 1981; Turcotte et€al., 1990; Woods, Melnik and Sparks, 2002); and (3) threshold vol-
1995). The decompression wave is followed ume fraction (Sparks, 1978).
closely by a fragmentation wave that may
travel more slowly through the bubbly magma 7.3.1â•… Strain-rate threshold
(Melnik and Sparks, 2002; Spieler et€ al., 2004a; High magma acceleration rates can generate
Fig. 7.1b). Behind the fragmentation front a mix- large strain rates that may result in a rheo-
ture of expanding gases and freshly produced logical transition from ductile to brittle behavior
pyroclasts is projected upward and expelled (Dingwell, 1996, 2001; Papale, 1999). Essentially,
from the conduit. The fragmentation wave is under high acceleration or high strain-rate
generally thought to fragment and quench the conditions, the magma relaxation or response
magma faster than dissolved gases can exsolve time may exceed the time over which stresses
in response to the decompression. Therefore, to are applied, causing the magma to behave in a
first order, exsolution of magmatic volatiles is brittle fashion, and leading to fragmentation.
assumed to be insignificant during fragmenta- Magma relaxation time τ increases with increas-
tion, and thus only volatiles that were already ing magma viscosity, therefore both increasing
in the vapor phase prior to plug disruption par- viscosity and increasing strain rate favor brit-
ticipate in the eruption (Sparks, 1978; Woods, tle fragmentation (Fig. 7.4). This concept can
1995; Melnik and Sparks, 2002; Clarke et€ al., be expressed mathematically via the Maxwell
2002a,b, 2007). Eruptive products therefore pre- relation
serve to some extent the pre-explosion state of
du 1 G
magma vesiculation. However, up to several = κ = κ ∞ (Papale, 1999) (7.3)
percent by volume bubble expansion can occur dz τ η
syn-explosively (i.e., during eruption), due to and
both bubble nucleation and growth (Melnik
and Sparks, 2002; Formenti and Druitt, 2003; du 1 G
= κ = κ ∞ (Gonnermann and Manga, 2003),
Giachetti et€al., 2010) and thus erupted clasts may dr τ η
be more vesicular than the pre-eruptive magma. (7.4)
Furthermore, because the velocity of the frag-
mentation wave greatly exceeds magma ascent where du/dz is the spatial velocity gradient (or
velocity (as the magma may have been stalled spatial acceleration) in the direction of flow
by the sealing plug), the magma can be assumed (equivalent to elongational strain rate), du/dr
to be stationary prior to fragmentation (Kieffer, is the spatial velocity gradient in the direction
or all of the pyroclasts are < 1 mm in diameter Solutions to this system of equations (7.8)–
(Woods, 1995). The resulting ascent of the pyro- (7.10) show that, in the isothermal limit (i.e., all
clastic mixture, assuming one-dimensional flow particles in thermal equilibrium with the gas;
through a conduit of constant cross-sectional f = 1), the corresponding velocity at the front
area, can be expressed in terms of the conserva- of the expanding pyroclastic mixture can be
tion of mass, momentum, and energy as follows approximated as follows (Woods, 1995):
(Turcotte et€al., 1990; Woods, 1995):
1 P
( nRTγ ) ln 0 ,
1/ 2
∂ρ b ∂ ( ρ b u ) uf ≅ (7.11)
+ = 0 (conservation of mass). (7.8) γ P
∂t ∂z
where R is the gas constant for water vapor, T is
The first term represents unsteadiness in the
the mixture temperature, P is the pressure of the
system, the second term is the inertial term
mixture at any given time (atmospheric pressure
(represented by the spatial gradient), and the
Pa upon complete decompression), and P0 is the
zero on the right-hand side indicates that no
initial (pre-explosion) pressure inside the con-
mass enters or leaves the system. Here, ρb is the
duit. The ratio of specific heats for the mixture,
bulk density of the expanding mixture, t is time,
γ, approaches 1 for typical pyroclastic mixtures
z is the vertical coordinate, and u is the vertical
(Wohletz, 2001) and can be approximated as
velocity of the mixture.
nR
γ = 1+ . (7.12)
∂u ∂u 1 ∂P c g n + c m (1 − n) f
+u =− −g−F
∂t ∂z ρ b ∂z (7.9)
Solutions for a wide range of initial gas mass
(conservation of momentum).
fractions (0.01–0.1), initial temperatures (600–
The first term on the left-hand side is the tem- 1400 K), and initial pressure ratios across the
poral acceleration of the flow, and the second plug (0–100) reveal interesting trends (some
is the remaining inertial term. The first term solutions are shown in Fig. 7.5; Woods, 1995).
on the right-hand side is the pressure gradi- Vent velocity increases with increasing pressure
ent (which drives the flow; flow moves in the ratio and increasing volatile mass fraction. Mass
direction of a negative pressure gradient), g is flux per unit cross-sectional area of the conduit
gravitational acceleration which retards flow, increases with increasing pressure ratio (due
and F (approximated as 0.01; Wilson and Head, to an increase in vent velocity) and decreases
1981) represents the frictional forces that also with increasing volatile mass fraction (due to a
retard the flow. Typically, gravity and friction decrease in mixture density).
can be ignored because, during the first several As stated above, the isothermal assump-
seconds of the expansion, these two terms are tion breaks down when a significant portion of
small relative to the pressure, acceleration, and the pyroclasts are > 1 mm, resulting in higher
inertial terms. values of γ. Assuming dynamic equilibrium, the
corresponding solution for flow-front velocity is
1 (Woods 1995)
Pd + ( c m f ( 1 − n ) + c g n ) dT = 0
ρb (7.10) 1/ 2 γ −1
(conservation of energy), 2 nRT γ P0 2 − 1 .
uf ≅ (7.13)
γ −1 γ P
where f represents the proportion of solids in
thermal equilibrium with the gas during decom- Mixture velocity decreases and mixture dens-
pression, n is the mass fraction of volatiles par- ity increases with increasing levels of thermal
ticipating in the explosion, cm is the specific heat disequilibrium (increasing proportion of large
of the magma (~1100 J kg−1 K−1), cg is the specific clasts; Fig. 7.5), pushing the system toward
heat of the volatile phase at constant volume gravitational collapse and formation of pyro-
(~1000 J kg−1 K−1). clastic density currents. These general trends
Explosion velocity, u, (m s–1) 350 Eq. (7.11) (iso) been used to calculate vent velocities (Clarke
300 et€ al., 2002a); multiphase solutions are reason-
250 ably consistent with the pseudogas/shock tube
200 relationships. Details of the shock-tube relations
Eq. (7.13) (non-iso)
150
are presented in Section 7.5 and the various
model solutions are compared in Figure 7.5.
100
Eq. (7.18) The vent flux associated with vulcanian erup-
50
tions is thought to be highly impulsive. The ini-
0 tial phase accelerates rapidly and quickly wanes.
0 20 40 60 80 100 120
Pressure ratio across plug, P0/Pa (or p4/p1) When the fragmentation wave reaches unfrag-
mentable magma, the vent flux decays to near
Figure 7.5╇╇ Calculated vent velocities for a range of initial zero, although sustained gas exhalations fol-
pressure ratios, using four methods of calculation. The lowing the main pulse have been documented
Woods (1995) isothermal solution (Eq. (7.11)) with 0.01 for several eruptions (Hoblitt et€al., 1996; Druitt
mass fraction volatiles at 1000 K (solid black line) significantly
et€al., 2002). Seismic (Druitt et€al., 2002) and com-
exceeds velocities calculated using shock-tube equations
putational evidence (Clarke et€al., 2002a) of the
for pseudogas mixtures (Eq. (7.18); dashed line, where fluid
properties in the high-pressure, pre-explosion conduit, impulsive nature of the vent flux is shown in
region 4, are calculated to be γ4 = 1.004 and c4 = 30 m s−1), Figure 7.6.
and numerical solutions to full Navier–Stokes, multiphase
equations (triangles, Clarke et€al., 2002a). The non-isothermal
solution (Eq. (7.13); Woods, 1995) for 0.01 mass fraction
volatiles, initial temperature of 1000 K, and 3% of particles 7.5╇ ╇ Shock waves
< 1 mm (dashed-dotted line) is very similar to the shock-tube
and multiphase solutions.
Leading shock waves are a consequence (or
propagation) of the pressure discontinuity
are consistent with earlier solutions presented between the high-pressure, gas-rich magma in
in Self et€al. (1979). Note that this effect is inde- the conduit and the atmosphere. They propagate
pendent of clast settling velocity, which is not ahead of and, according to shock-tube theory,
considered here. In this regard, it is important drive motion of the pyroclastic mixture. Leading
to keep in mind that in reality large clasts may shock waves are sometimes visible because they
settle out of the rising mixture, contributing to condense atmospheric water vapor, allowing
the formation of pyroclastic density currents their properties to be documented and meas-
and/or reducing the density of the remaining ured (Nairn, 1976; Ishihara, 1985; Morrissey and
mixture, allowing it to rise buoyantly. Chouet, 1997). Shock waves are characterized
Other approaches (e.g., Ishihara, 1985; by a sharp increase in atmospheric pressure,
Chojnicki et€al., 2006) treat the system as a shock followed by a dip to a pressure that is less than
tube and calculate the velocity of the expanding ambient, producing a characteristic N-shaped
mixture according to equations for shock-driven wave, which has been documented by stationary
flow (Saad, 1985). In some cases, the mixture pressure sensors (Morrissey and Chouet, 1997).
is treated as a thermally and dynamically per- The amplitude of the wave decays nonlinearly
fectly coupled fluid, termed a pseudogas. The with decreasing initial pressure ratio across the
sound speed c and ratio of specific heats γ of a plug and with increasing distance from source.
pseudogas are significantly less than values for For the simple case of adiabatic and inviscid
the gas alone (Kieffer, 1981; Dobran et€al., 1993; flow of an ideal gas, shock characteristics have
Wohletz, 2001). Shock-tube/pseudogas methods been derived in terms of the pre-explosion pres-
for a given pre-explosion pressure ratio tend to sure ratio across the plug (p4/p1) by solving the
predict lower vent velocities relative to solu- conservation of mass, momentum, and energy
tions of Turcotte et€al. (1990) and Woods (1995). equations over a control volume that encom-
Multiphase models that account for heat and passes and travels along with the shock wave
momentum exchange between phases have also (Saad, 1985). In Equations (7.14)−(7.18), subscript
–2
–4
0 10 20 30 40 50 60 70
Time, t (s)
(d)
7
Mass flux ( 107 kg s–1)
0
Time, t (s)
1 represents atmospheric conditions, 2 repre- ratio across the plug p4/p1 can be expressed in
sents conditions in the region behind the shock, terms of shock strength
but outside the conduit, and 4 represents prop- 2γ 4
erties in the pre-explosion, high-pressure con-
−
(γ 4 − 1)
c p
duit (Regions 1, 2 and 4 in Fig. 7.1). Assuming ( γ 4 − 1) c 1 p2 − 1
p 4 p2 4 1
adiabatic and inviscid flow, the shock strength, = 1− ,
or pressure ratio across the shock p2/p1 (Fig. 7.1), p1 p1 p
2γ 1 2γ 1 + ( γ 1 + 1) 2 − 1
can be expressed in terms of the shock Mach p1
number, Ms = vs/c, where vs is the shock velocity (7.15)
and c is the speed of sound in the atmosphere
(~330 m s−1 for a standard terrestrial atmos- and in terms of Ms
phere; Saad, 1985)
2γ 4
−
p2 2γ 1 γ 4 − 1 c1 2 (γ 4 − 1)
= 1+
γ1 + 1
( M2s − 1) , (7.14) (
Ms − 1 )
p1
p4 = γ 1 − 1 2γ 1 M 2 − 1 1 − γ 4 + 1 c 4 .
p1
γ1 + 1 γ1 −1
s
Ms
where γ1 is the ratio of specific heats for the fluid
into which the shock is propagating, in this case,
(7.16)
the atmosphere. In turn, the initial pressure
RT ρg
c4 = n + (1 − n) , (7.17)
n ρm
magma, the resulting pyroclastic mixture may Reynolds number, defined as the ratio of
enter the atmosphere in an overpressured state. inertial to viscous forces, can be written for vul-
This means that compressibility may be critical canian jets as
to understanding the subsequent propagation
ρ b uL
of the pyroclastic mixture. The combination Re = , (7.19)
ηb
of field observations from the Soufrière Hills
volcano, Montserrat, British West Indies, and where u is flow velocity, L is a characteristic
multi-phase models for gas–particle mixtures length scale (usually the vent diameter), and ηb
(similar to that used in Morrissey and Chouet, is the bulk dynamic viscosity of the pyroclastic
1997) shows that pressure inside the volcanic mixture. Re values for typical vulcanian erup-
jets adjusts to atmospheric pressure within 50 tions range from 105 to 107 and even as high
m from the vent (less than two vent diameters) as 108 for some large diameter jets; u can be as
and that initially supersonic flows decelerate high as 400 m s−1; L is on the order of tens of
to subsonic conditions over similar distances meters, although some vents can be hundreds of
(Formenti et€al., 2003). This finding is reasonably meters in diameter; and ρb can range from less
consistent with other field observations (Hoblitt, than 1 kg m−3 (less than atmospheric density,
1986, Sparks et€ al., 1997; Clarke et€ al., 2002a; and therefore buoyant) to as high as 100 kg€m−3;
Druitt et€al., 2002; Formenti et€al., 2003). We can corresponding values of ηb are 10−4−10−1 Pa s
apply, with some assumptions, relationships (Wohletz, 2001). Although some characteristics
derived from experiments on steady, super- such as the length scales of the smallest turbu-
sonic, overpressured, ideal-gas jets (Yüceil and lent eddies depend on Re, nearly all vulcanian
Ötügen, 2002) to the problem of vulcanian erup- eruptions are well within the turbulent regime,
tions. For initial conduit pressures ranging from and thus inertial forces far outweigh viscous
1 to 10 MPa, fluid properties of volcanic gas–par- forces, and in general flow can be assumed to be
ticle mixtures, and conduit diameters <€100 m, inviscid, except near boundaries.
we find that vulcanian jets should reach atmos- Richardson number (Ri) is defined as the
pheric pressure by no more than 700 m from ratio of buoyancy forces to inertial forces and
source in the most extreme cases, and by no can be expressed as
more than 100 m from source in typical erup- gL
tions. During the decompression to atmospheric Ri = , (7.20)
u2
pressure, the jet expands in width by a factor of
where the reduced gravity, g′ = g(ρb − ρa)/ρa, and
2−8. Over this same distance, the jet accelerates,
ρa is the ambient (atmospheric) density at the
but only by as much as 10%. Simultaneously, the
vent. Ri expresses whether a vulcanian eruption
density of the mixture decreases significantly,
is dominated by positive or negative buoyancy
by up to a factor of 70.
forces (large Ri), by momentum forces (small
Ri), or whether both momentum and buoyancy
are important (Ri ~ 1).
7.7╇ ╇Pyroclastic phase: vertical Total buoyancy injected B can be expressed as
ascent B = g′â•›V, where V is the volume of fluid injected
(time integrated volume flux), and total momen-
Key fluid dynamics principles that should be tum injected M can be expressed as M = u0V,
considered when describing the subsequent where u0 is the initial velocity at the vent.
pyroclastic phase (beyond the vent) of vul- Ascent of the pyroclastic mixture associated
canian eruptions are Reynolds number (Re), with a vulcanian eruption may be described by
reduced gravity (g’), Richardson number (Ri), one of three simplified models: they may behave
total injected buoyancy (B), and total injected as thermals (short releases of a buoyant fluid),
momentum (M). impulsive jets (short injections of fluid imparting
(0.01) for individual jets, and this inefficient typical volcanic conditions (Fagents and Wilson,
entrainment is thought to have contributed to 1993; Waitt et€al., 1995; Mastin, 2001).
collapse and pyroclastic flow formation. In the vertical (z) direction, the ballistic block
is subject to both drag and gravitational forces:
dv z − v z ρa v′ACd ρ − ρa
7.8╇ ╇ Ballistic analysis
dt
=
2m
−g r
ρr
. (7.26)
Ballistic blocks or bombs are typically associ- The second term on the right accounts for the
ated with vulcanian eruptions and are thought buoyancy of the clast in air; however, in most
to represent the disrupted sealing plug. Ballistic cases, ballistic block density ρr greatly exceeds
block fields have been documented carefully for atmospheric density and the second term goes
the 1977 phreatomagmatic eruptions of Ukinrek to –g. Vertical velocity can be expressed as
Maars, Alaska (Self et€al., 1980), the 1992 subplin- v sin θ.
ian eruptions of Crater Peak Vent, Mount Spurr Solutions to Eqs. (7.25) and (7.26) can be
volcano, Alaska (Waitt et€al., 1995), the 1997 vul- used to interpret observed ballistic trajectories
canian eruptions of the Soufrière Hills volcano (range, size, and perhaps launch angle if avail-
(Druitt et€al., 2002), and the 1999 vulcanian erup- able) in terms of initial vent velocity (Wilson,
tions of Guagua Pichincha volcano, Ecuador 1972; Fagents and Wilson, 1993; Waitt et€ al.,
(Wright et€ al., 2007), among others. Blocks on 1995; Mastin, 2001) and sometimes in terms of
the order of a half meter in diameter can be initial vent pressure (Ishihara, 1985). Because
launched to 3 km from the vent, and in some of the high velocity of the ejecting mixture sur-
cases smaller blocks can reach > 6 km from the rounding the ballistic, one may ignore drag in
vent. In general, ballistic size decreases with dis- the region immediately around the vent because
tance from the vent. However, in some cases the the ejecting fluid is moving at (nearly) the same
opposite is true because of complex drag inter- velocity as the ballistic clast (v′ = 0) (Fagents and
actions between the blocks and the expanding Wilson, 1993); or in some cases, the accelerat-
pyroclastic mixture and the surrounding air (Self ing effect of a surrounding pyroclastic mixture
et€al., 1980; Waitt et€al., 1995; de’ Michieli Vitturi must be considered (de’ Michieli Vitturi et€ al.,
et€al., 2010). Their sizes, trajectories, ranges, and 2010). Overestimation of the drag coefficient,
textures can be used to infer pre-explosion or total drag near the vent, or launch angle results
vent conditions. in a significant overestimate of initial ballistic
Ballistic trajectories can be calculated by velocity.
solving Newton’s second law of motion, F =
ma (force equals mass times acceleration) for a
known launch velocity and angle. In the hori-
7.9╇ ╇Transitions in eruption style or
zontal (x) direction, the ballistic clast is subject
to drag forces alone: scale
dv x − v x ρ a v′ACd Vulcanian eruptions may transform into sus-
= . (7.25)
dt 2m tained, quasi-steady explosive eruptions or
Horizontal velocity vx is equivalent to v cos θ, may end suddenly to be replaced by effusive
where v is the total velocity of the clast and θ dome-building. As stated in Section 7.3, it is rea-
is the trajectory angle relative to the horizon- sonable to assume that gas does not diffuse from
tal; A is the cross-sectional area of the clast; v′ the melt into bubbles during propagation of the
is the clast velocity relative to the motion of the fragmentation front into the conduit. However,
surrounding fluid (= v€ – w); and Cd is the drag when volatile diffusion is fast and efficient, such
coefficient (a function of particle Re, where clast as for high bubble number densities and large
diameter is the characteristic length scale in diffusion coefficients, this assumption becomes
Eq. (7.19)) and may vary between 0.06 and 1 for invalid. For these cases, numerical solutions to
complex systems of equations have been used 2006). Corresponding solutions show that no
to explore the effects of syn-explosion gas dif- gas exsolution during fragmentation results
fusion on eruption characteristics (Melnik and in a single explosive pulse (until magma can
Sparks, 2002; Mason et€al., 2006). Results suggest slowly ascend in response to conduit evacu-
that indeed some amount of syn-explosion gas ation and prepare for another explosion), inter-
diffusion may extend the scale and duration of a mediate (disequilibrium) syn-fragmentation gas
given vulcanian eruption, may explain repeated, exsolution leads to pulsatory eruptions, and the
pulsatory explosions, and may result in a transi- equilibrium case results in sustained eruption
tion in eruptive style, leading to a quasi-steady (Melnik and Sparks, 2002). Greater total volatile
subplinian eruption. content for a given set of conditions pushes the
The role of syn-fragmentation gas exsolu- system toward multiple pulses or quasi-steady
tion can be determined by examining the Peclet behavior (Mason et€ al., 2006). The addition of
number (Pe), which is the ratio of the timescale crystals tends to increase the depth to which the
for gas exsolution to the timescale of fragmenta- fragmentation front reaches for a single pulse
tion (Navon and Lyakhovsky, 1998; Melnik and because of increased viscosity. However mag-
Sparks, 2002; Proussevitch and Sahagian, 2005; mas with low crystal fractions (low viscosities)
Koyaguchi and Mitani, 2005; Mason et€al., 2006). are more likely to stabilize into a steady-state
The timescale of exsolution is determined by eruption because the magma will more eas-
the rate of diffusion of a volatile in the melt of ily ascend to meet and feed the fragmentation
interest and the length scale in question. The wave (Mason et€al., 2006).
timescale of fragmentation is a function of the Transition from periodic explosions to effu-
distance between the decompression wave and sive activity may occur when permeability,
the fragmentation front and the velocity of the rather than fragmentation, develops either
fragmentation front relative to the (possibly) throughout the magma (Mueller et€al., 2008) or
ascending unfragmented magma; this relative along conduit walls (Gonnermann and Manga,
velocity is simply the difference between the 2003; Tuffen et€ al., 2003), and therefore sup-
two velocity vectors (Melnik and Sparks, 2002). presses explosion. A particular sequence of peri-
When Pe ≫ 1, syn-fragmentation gas exsolution odic explosive eruptions at the Soufrière Hills
can be ignored; when Pe ≪ 1, it must be consid- volcano in 2003 is thought to have ended in this
ered. The degree of gas exsolution may be quan- way via permeable escape of gas along conduit
tified according to a diffusion parameter that walls (Edmonds and Herd, 2007).
describes the intensity of gas exsolution dur-
ing fragmentation; this parameter is a function
of bubble number density per unit volume of
magma and the diffusion coefficient of water in 7.10╇ ╇Summary and future
the melt (Melnik and Sparks, 2002; Mason et€al., directions
2006). End-members are zero gas exsolution
(the simplifying assumption in Section 7.3) and Vulcanian eruptions are very common and may
the equilibrium case in which gas exsolution present significant hazards to surrounding
instantaneously responds to decompression. populations. Their impulsive nature, unsteady
The system can be represented by dynamics, as well as the potential importance
one-dimensional conservation equations that of compressible fluid flow, makes theoretical
account for the difference in pressure between understanding of their characteristics difficult.
the bubbles and the magma (Eq. (7.2)), allow Furthermore, factors contributing to setting up
gas exsolution during fragmentation (depend- conditions sufficient for explosion initiation
ing on Pe), and allow underlying magma to remain less than fully understood, primarily
ascend in response to the decompression asso- because complex relationships exist among
ciated with conduit evacuation (Melnik and magma degassing and crystallization, magma
Sparks 2002; Melnik et€ al., 2005; Mason et€ al., permeability, and pressurization of the shallow
conduit. Vulcanian eruptions may transform into • Leading shock waves may propagate through
sustained explosive volcanic eruptions, or may the atmosphere ahead of the eruption front.
be replaced by effusion, offering an interesting • Ejection of ballistic blocks, which represent
focus of study for those interested in transitions the former conduit plug, often characterizes
in eruption style. Future observations and meas- the initial phases of vulcanian eruptions.
urements of dynamic conditions will improve • Quantitative characteristics of shock waves,
understanding. Useful measurements include explosion flow front, and ballistic blocks and
temporal evolution of plume and pyroclastic their trajectories can be used to constrain vent
flow morphology and flow-front velocity using fluxes and pre-eruption conduit conditions.
simple stationary video camera (e.g., Sparks
and Wilson, 1982). These values can be com-
pared to theoretical understanding of controls 7.12╇ ╇ Notation
on flow propagation (thermal, impulsive jet,
forced thermal), and be used to calculate total a constant equal to 8.21 x 105 (MPa m−1)
mass erupted (Clarke et€ al., 2009) and entrain- A cross-sectional area of ballistic block (m2)
ment coefficients (Formenti et€al., 2003). Profiles B total buoyancy injected/erupted, = g′V
of plume and pyroclastic flow temperature, vel- (m4 s−2)
ocity, and particle concentration are also useful c speed of sound in the atmosphere
data. Currently such measurements are very c1 speed of sound in fluid in low-pressure
difficult or impossible to make; however new region (m s−1)
ground-based imaging techniques (e.g., Doppler c4 speed of sound in fluid in high-pressure
radar, thermal-IR, and UV cameras; Yamamoto region (m s−1)
et€ al., 2008) may make them possible in the cm specific heat of the magma
future, offering a potentially fruitful pathway (J kg−1 K−1)
for development and testing of new measure- cg specific heat of the volatile phase
ment techniques. (J kg−1 K−1)
Cd drag coefficient for ballistic block
d diameter of plume/jet at the vent (m)
7.11╇ ╇ Summary D conduit diameter (m)
f the proportion of solids in thermal
• Vulcanian eruptions are characteristically equilibrium with the gas during eruption
short-lived and are fed by highly unsteady F frictional forces per unit mass (N kg−1)
vent flux, distinguishing them from subplin- g gravitational acceleration (m s−2)
ian eruptions. g′ reduced gravity, = g(ρb– ρa)/ρa (m s−2)
• Vulcanian eruptions represent a transitional G∞ elastic modulus of the magma (Pa)
eruption style, and may occur before or after k permeability of a bubbly magma (m2)
long-lived explosive eruptions or extended m mass of ballistic block (kg)
periods of dome growth. L characteristic length scale (m)
• Pyroclast grain size and dispersal area for vulca- M total momentum injected/erupted,
nian eruptions do not follow a particular trend; = vz0V (m4 s−1)
vulcanian deposit characteristics overlap with Mass total mass erupted (kg)
many other styles of explosive eruption. n mass fraction of volatiles in the erupting
• Plug formation and disruption are critical to mixture
generating vulcanian eruptions. N Brunt–Väisälä (buoyancy) frequency (s−1)
• During the main phase of a vulcanian erup- p2/p1 shock strength, ratio of pressures behind
tion, a fragmentation front progressively and ahead of shock
accesses deeper and deeper magma, which p4/p1 ratio of pressures in high-pressure
in part explains the varied nature of eruption (pre-explosion conduit) region to
products. low-pressure region above plug
Cas, R. A. F. and Wright, J. V. (1987). Volcanic Dingwell, D. B. (1996). Volcanic dilemma: flow or
Successions: Modern and Ancient. London: Unwin blow? Science, 273, 1054–1055.
Hyman Ltd. Dingwell, D. B. (2001). Magma degassing and
Cashman, K. and Blundy, J. (2000). Degassing and fragmentation: Recent experimental advances. In
crystallization of ascending andesite and dacite. From Magma to Tephra, ed. A. Freundt and M. Rosi.
Philosophical Transactions of the Royal Society, 358, Amsterdam: Elsevier, pp. 1–23.
1487–1513. Dobran, F., Neri, A. and Macedonio, G. (1993). Numerical
Cashman, K. V. and McConnell, S. M. (2005). Multiple simulation of collapsing volcanic volumns. Journal
levels of magma storage during the 1980 summer of Geophysical Research, 98, 4231–4259.
eruptions of Mount St. Helens, WA. Bulletin of D’Oriano, C., Poggianti, E., Bertagnini, A. et€al. (2005).
Volcanology, 68, 57–75. Changes in eruptive style during the A.D. 1538
Chojnicki, K., Clarke, A. B. and Phillips, J. C. (2006). Monte Nuovo eruption (Phlegrean Fields, Italy):
A shock-tube investigation of the dynamics of The role of syn-eruptive crystallization. Bulletin of
gas–particle mixtures: Implications for explosive Volcanology, 67, 601–621.
volcanic eruptions. Geophysical Research Letters, 33, Druitt, T. H., Young, S. R., Baptie, B. et€al. (2002).
L15309, doi:10.1029/2006FL026414. Episodes of cyclic Vulcanian explosive activity
Clarke, A. B., Neri, A., Macedonio, G., Voight, B. and with fountain collapse at Soufrière Hills volcano,
Druitt, T. H. (2002a). Computational modeling of Montserrat. In The Eruption of Soufrière Hills Volcano,
the transient dynamics of August 1997 Vulcanian Monsterrat, from 1995 to 1999, ed. T. H. Druitt and
explosions at Soufrière Hills volcano, Montserrat: B.€P. Kokelaar. Geological Society, London, Memoir
Influence of initial conduit conditions on near-vent 21, pp. 281–306.
pyroclastic dispersal. In The Eruption of Soufrière Hills Dufek, J. and Bergantz, G. W. (2005). Transient
Volcano, Monsterrat, from 1995 to 1999, ed. T. H. Druitt two-dimensional dynamics in the upper conduit
and B. P. Kokelaar. Geological Society, London, of a rhyolitic eruption: A comparison of closure
Memoir 21, pp. 319–348 models for the granular stress. Journal of Volcanology
Clarke, A. B., Neri, A., Macedonio, G. and Voight,€B. and Geothermal Research, 143, 113–132.
(2002b). Transient dynamics of Vulcanian Edmonds, M. and Herd, R. A. (2007). A volcanic
explosions and column collapse. Nature, 415, degassing event at the explosive-effusive
897–901. transition. Geophysical Research Letters, 34, L21310,
Clarke, A. B., Stephens, S., Teasdale, R., Sparks,€R.€S.€J. doi:10.1029/2007GL031379.
and Diller, K. (2007). Petrological constraints Eichelberger, J. C., Carrigan, C. R., Westrich, H.€R.
on the decompression history of magma prior and Price, R. H. (1986). Non-explosive silicic
to Vulcanian explosions at the Soufrière Hills volcanism. Nature, 323, 598–602.
volcano, Montserrat. Journal of Volcanology and Fagents, S. A. and Wilson, L. (1993). Explosive
Geothermal Research, 161, 261–274. volcanic eruptions€– VII. The ranges of pyroclasts
Clarke, A. B., J. C. Phillips, and K. N. Chojnicki (2009). ejected in transient volcanic explosions. Geophysical
An Investigation of the Dynamics of Vulcanian Journal International, 113, 359–370.
Eruptions using Laboratory Analogue Experiments Formenti, Y. and Druitt, T. H. (2003). Vesicle
and Scaling Analysis, in Studies in Volcanology: The connectivity in pyroclasts and implications for the
Legacy of George Walker, edited by T. Thordarson fluidization of fountain-collapse pyroclastic flows,
et€al., pp. 416, Geological Society, London, IAVCEI Montserrat (West Indies). Earth and Planetary Science
Publications, 2. Letters, 214, 561–574.
de’ Michieli Vitturi, M., Neri, A., Esposti Ongaro,€T. Formenti, Y., Druitt, T. H. and Kelfoun, K. (2003).
Lo Savio, S. and Boschi, E. (2010). Lagrangian Characterization of the 1997 Vulcanian explosions
modeling of large volcanic particles: Application to of Soufrière Hills Volcano, Montserrat, by video
Vulcanian explosions. Journal of Geophysical Research, analysis, I. Bulletin of Volcanology, 65, 587–605.
115, B08206, doi:10.1029/2009JB007111. Fox, R. W., McDonald, A. T. and Pritchard, P. J. (2006),
Diller, K. D., A. B. Clarke, B. Voight, and A. Introduction to Fluid Mechanics, 6th edn. Hoboken, NJ:
Neri (2006). Mechanisms for Conduit Plug John Wiley and Sons.
Formation: implications for Vulcanian Geschwind, C. H. and Rutherford, M. J. (1995).
explosions. Geophysical Research Letters, 33, L20302, Crystallization of microlites during magma ascent:
doi:10.1029/2006GL027391. the fluid mechanics of 1980–1986 eruptions
at Mount St. Helens. Bulletin of Volcanology, 57, study at high viscosities. Journal of Geophysical
356–370. Research, 100(3), 4215–4229.
Giachetti, T., Druitt, T. H., Burgisser, A., Arbaret, L. Mason, R. M., Starostin, A. B., Melnik, O. E. and
and Galven, C. (2010). Bubble nucleation, growth Sparks, R. S. J. (2006). From Vulcanian explosions
and coalescence during the 1997 Vulcanian to sustained explosive eruptions: the role of
explosions of Soufrière Hills Volcano, Montserrat. diffusive mass transfer in conduit flow dynamics.
Journal of Volcanology and Geothermal Research, Journal of Volcanology and Geothermal Research, 153,
193(3–4), 215–231. 148–165.
Glaze, L. S., Francis, P., Self, S. and Rothery, D. A. Mastin, L. G. (2001). A simple calculator of ballistic
(1989). The 16 September 1986 eruption of Lascar trajectories for blocks ejected during volcanic
volcano, north Chile: satellite investigations. eruptions. United States Geological Survey Open-File
Bulletin of Volcanology, 51, 149–160. Report 01–45, 16pp., available at http://vulcan.
Gonnermann, H. M. and Manga, M. (2003). Explosive wr.usgs.gov/Projects/Mastin/Publications/OFR01–45/
volcanism may not be an inevitable consequence framework.html.
of magma fragmentation. Nature, 426, 432–435, McBirney, A. R. and Murase, T. (1971). Factors
doi:10.1038/nature02138. governing the formation of pyroclastic rocks,
Hammer, J. E., Cashman, K. V., Hoblitt, R. P. and Bulletin Volcanologique, 34(2), 372–384.
Newman, S. (1999). Degassing and microlite Melnik, O. and Sparks, R. S. J. (1999). Nonlinear
crystallization during pre-climactic events of the dynamics of lava dome extrusion. Nature, 402, 37–41.
1991 eruption of Mt. Pinatubo, Philippines. Bulletin Melnik, O. and Sparks, R. S. J. (2002). Modelling of
of Volcanology, 60, 355–380. conduit flow dynamics during explosive activity at
Hess, K. U. and Dingwell, D. B. (1996). Viscosities Soufriere Hills Volcano, Montserrat. In The Eruption
of hydrous leucogranitic melts: A non-Arrhenian of Soufrière Hills Volcano, Monsterrat, from 1995 to
model. American Mineralogist, 81, 1297–1300. 1999, ed. T. H. Druitt and B. P. Kokelaar. Geological
Hoblitt, R. (1986). Observations of the Eruptions of July 22 Society, London, Memoir 21, pp. 307–317.
and August 7, 1980, at Mount St. Helens, Washington. Melnik, O., Barmin, A. A. and Sparks, R. S. J. (2005).
United States Geological Survey Professional Paper Dynamics of magma flow inside volcanic conduits
1335, 44 pp. with bubble overpressure buildup and gas loss
Hoblitt, R. P., Wolfe, E. W., Scott, W. E. et€al. (1996). through permeable magma. Journal of Volcanology
The preclimactic eruptions of Mount Pinatubo, and Geothermal Research, 143, 53–68.
June 1991. In Fire and mud: Eruptions and lahars of Mercalli, G. (1907). I Volcani Attivi della Terra. Milano:
Mount Pinatubo, Philippines, ed. C. G. Newhall and Ulrico Hoepli.
R. S. Punongbayan. Seattle, WA: University of Moore, G. and Carmichael, I. S. E. (1998). The
Washington Press, pp. 457–511. hydrous phase equilibria (to 3 kbar) of an andesite
Ishihara, K. (1985). Dynamical analysis of volcanic and basaltic andesite from western Mexico:
explosion. Journal of Geodynamics, 3, 327–349. constraints on water content and conditions of
Kaminski, E., Tait, S. and Carazzo, G. (2005). phenocryst growth. Contributions to Mineralogy and
Turbulent entrainment into jets with Petrology, 130, 304–319.
arbitrary buoyancy. Journal of Fluid Mechanics, Morrissey, M. M. and Chouet, B. A. (1997). Burst
526, 361–376. conditions of explosive volcanic eruptions
Koyaguchi, T. (2005). An analytical study for recorded on microbarographs. Science, 275,
1-dimensional steady flow in volcanic conduits. 1290–1293.
Journal of Volcanology and Geothermal Research, 143, Morrissey, M. M. and Mastin, L. G. (2000). Vulcanian
29–52. Eruptions. In Encyclopedia of Volcanoes, ed. H.
Koyaguchi T. and Mitani, N. K. (2005). A theoretical Sigurdsson. San Diego, CA: Academic Press, pp.
model for fragmentation of viscous bubbly 463–475.
magmas in shock tubes. Journal of Geophysical Morton, B. R. (1959). Forced plumes. Journal of Fluid
Research, 110, B10202, doi:10.1029/2004JB003513. Mechanics, 5, 151–163.
Kieffer, S. W. (1981). Blast dynamics at Mount St. Mueller, S., Scheu, B., Spieler, O. and Dingwell,€D.€B.
Helens on 18 May 1980. Nature, 291, 568–570. (2008). Permeability control on magma
Lejeune, A.-M. and Richet, P. (1995). Rheology of fragmentation. Geology, 36(5), 399–402.
crystal-bearing silicate melts; an experimental doi:10.1130/G24605A.
Nairn, I. A. (1976). Atmospheric shock waves and volcanic jets. Journal of Geophysical Research, 114,
condensation clouds from Ngaurhoe explosive B10203, doi:10.1029/2009JB006298.
eruptions. Nature, 259, 190–192. Sparks, R. S. J. (1978). The dynamics of bubble
Nairn, I. A. and Self, S. (1978). Explosive eruptions formation and growth in magmas: a review
and pyroclastic avalanches from Ngauruhoe in and analysis. Journal of Volcanology and Geothermal
February 1975. Journal of Volcanology and Geothermal Research, 3, 1–37.
Research, 3, 39–60. Sparks, R. S. J. (1997). Causes and consequences of
Navon, O. and Lyakhovsky, V. (1998). Vesiculation pressurization in lave dome eruptions. Earth and
processes in silicic magmas. In The Physics of Planetary Science Letters, 150, 177–189.
Explosive Volcanic Eruptions, ed. J. S. Gilbert and Sparks, R. S. J. and Wilson, L. (1982). Explosive
R. S. J. Sparks. Cambridge, UK: Cambridge volcanic eruptions€– V. Observations of plume
University Press, pp. 27–50. dynamics during the 1979 Soufrière eruptions, St.
Papale, P. (1999). Strain-induced magma Vincent. Geophysical Journal of the Royal Astronomical
fragmentation in explosive eruptions. Nature, 397, Society, 69, 551–570.
425–428. Sparks, R. S. J., Bursik, M. I., Carey, S. N. et€al. (1997).
Proussevitch, A. A. and Sahagian, D. L. (1996). Volcanic Plumes. New York, NY: John Wiley and Sons.
Dynamics of coupled diffusive and decompressive Spieler, O., Dingwell, D. B. and Alidibirov, M. (2004a).
bubble growth in magmatic systems. Journal of Magma fragmentation speed: an experimental
Geophysical Research, 101, 17,156–17,447. determination. Journal of Volcanology and Geothermal
Proussevitch, A. and Sahagian, D. (2005). Research, 129, 109–123.
Bubbledrive-1: A numerical model of volcanic Spieler, O., Kennedy, B., Kueppers, U. et€al. (2004b).
eruption mechanisms driven by disequilibrium The fragmentation threshold of pyroclastic rocks.
magma degassing. Journal of Volcanology and Earth and Planetary Science Letters, 226, 139–148.
Geothermal Research, 143, 89–111. Stix, J., Torres, R. C., Narváez M., L. et€al. (1997). A
Rose, W. I. (1980). Gas and hydrogen isotopic model of Vulcanian eruptions at Galeras volcano,
analyses of volcanic eruptions clouds in Guatemala Colombia. Journal of Volcanology and Geothermal
sampled by aircraft. Journal of Volcanology and Research, 77, 285–303.
Geothermal Research, 7, 1–10. Taddeucci, J., Pompilio, M. and Scarlato, P. (2004).
Rose, W. I., Anderson, A. T., Woodruff, L. G. and Conduit processes during the July–August 2001
Bonis, S. B. (1978). The October 1974 basaltic explosive activity of Mt. Etna (Italy): Inferences
tephra from Fuego Volcano: description and from glass chemistry crystal size distribution of
history of the magma body. Journal of Volcanology ash particles. Journal of Volcanology and Geothermal
and Geothermal Research, 4, 3–53. Research, 137, 33–54.
Saad, M. A. (1985). Compressible Fluid Flow. Upper Tait, S. R., Jaupart, C. and Vergniolle, S. (1989).
Saddle River, NJ: Prentice-Hall, Inc. Pressure, gas content and eruptive periodicity of
Schmincke, H.-U. (1977). Eifel-Vulkanismus a shallow crystallizing magma chamber. Earth and
östlich des Gebiets Rieden-Mayen. Fortschritte der Planetary Science Letters, 92, 107–123.
Mineralogie, 55, 1–31. Takeuchi, S., Nakashima, S. Tomiya, A. and
Scott, W. E., Hoblitt, R. P., Torres, R. C. et€al. (1996). Shinohara, H. (2005). Experimental constraints
Pyroclastic flows of the June 15, 1991, climactic on the low gas permeability of vesicular magma
eruption of Mount Pinatubo. In Fire and Mud: during decompression. Geophysical Research Letters,
Eruptions and Lahars of Mount Pinatubo, Philippines, ed. 32, L10312, doi:10.1029/2005GL022491.
C. G. Newhall and R. S. Punongbayan. Seattle, WA: Tuffen, H., Dingwell, D. B. and Pinkerton, H. (2003).
University of Washington Press, pp. 545–570. Repeated fracture and healing of silicic magma
Self, S., Wilson, L. and Nairn, I. A. (1979). Vulcanian generate flow banding and earthquakes? Geology,
eruption mechanisms. Nature, 277, 440–443. 31(12), 1089–1092.
Self, S., Kienle, J. and Huot, J. -P. (1980). Ukinrek Turcotte, D. L., Ockendon, H., Ockendon, J. R. and
Maars, Alaska, II. Deposits and formation of the Cowley, S. J. (1990). A mathematical model of
1977 craters. Journal of Volcanology and Geothermal vulcanian eruptions. Geophysical Journal International,
Research, 7, 39–65. 103, 211–217.
Solovitz, S. A. and Mastin, L. A. (2009). Experimental Turner, J. S. (1973). Buoyancy Effects in Fluids. New York:
study of near-field air entrainment by subsonic Cambridge University Press.
Voight, B., Sparks, R. S. J., Miller, A. D. et€al. (1999). an ultraviolet imaging camera. Geophysical Research
Magma flow instability and cyclic activity at Letters, 35, L08314, doi:10.1029/2007GL032008.
Soufrière Hills Volcano, Montserrat. Science, Yüceil, K. B. and Ötügen, M. V. (2002). Scaling
283(5405), 1138–1142. parameters for underexpanded supersonic jets,
Waitt, R. B., Mastin, L. G. and Miller, T. P. (1995). Physics of Fluids, 14, 4206–4215.
Ballistic showers during Crater Peak eruptions of Zhang, Y. (1998). Experimental simulations of
Mount Spurr Volcano, summer 1992. United States gas-driven eruptions: kinetics of bubble growth
Geological Survey Bulletin, 2139, 89–106. and effect of geometry. Bulletin of Volcanology 59,
Walker, G. P. L. (1973). Explosive volcanic eruptions€– 281–290.
a new classification scheme. Geologische Rundschau, Zhang, Y. (1999). A criterion for the fragmentation
62(2), 431–446. of bubbly magma based on brittle failure theory.
Wilson, L. (1972). Explosive volcanic eruptions€– Nature, 402, 648–650.
II. The atmospheric trajectories of pyroclasts.
Geophysical Journal of the Royal Astronomical Society,
30, 381–392. Exercises
Wilson, L. and Head, J. (1981). Ascent and erupton
of magma on the Earth and moon. Journal of 7.1 Given the shock velocity in air of 450 m s−1 associ-
Geophysical Research, 86, 2971–3001. ated with the eruption of Sakurajima volcano in
Wilson, L., Sparks, R. S. J., Huang, T. C. and Watkins, 1982 (Ishihara, 1985), calculate:
N. D. (1978). The control of volcanic eruption column (a) The initial pressure ratio across the plug by
heights by eruption energetics and dynamics. Journal assuming that the mixture behaves as a pseu-
of Geophysical Research, 83, 1829–1836. dogas with c4 = 60 m s−1, γ4 = 1.04, c1 = 330
Wilson, L., Sparks, R. S. J. and Walker, G. P. L. (1980). m€s−1, γ1 = 1.4.
Explosive volcanic eruptions€– IV. The control (b) The shock velocity and strength for an ini-
of magma properties and conduit geometry on tial pressure ratio of 50, given the same fluid
eruption column behaviour. Geophysical Journal of properties.
the Royal Astronomical Society, 63, 117–148.
Wohletz, K. H. (2001). Pyroclastic surges and 7.2 Given the values used in and calculated for the
two-phase compressible flows. In From Magma to Sakurajima case above:
Tephra, ed. A. Freundt and M. Rosi. Amsterdam: (a) Calculate the corresponding vent velocity
Elsevier, pp. 247–312. using the isothermal relationship of Woods
Wohletz K. H. and Valentine, G. A. (1990). Computer (1995) assuming T = 1200 K, γ = 1.04, R = 460 J
simulations of explosive volcanic eruptions. kg−1 K−1, and n = 0.03.
In Magma Transport and Storage, ed. M. P. Ryan. (b) Calculate the vent velocity using shock tube
London: John Wiley and Sons, pp. 113–135. relationships.
Woods, A. W. (1995). A model of vulcanian (c) Calculate vent velocity using the non-isother-
explosions. Nuclear Engineering and Design, mal relationship of Woods (1995) assuming
155, 345–357. that the gas temperature drops to 500 K upon
Wright, H. M. N., Cashman, K. V., Rosi, M. and decompression with corresponding γ = 1.3
Cioni, R. (2007). Breadcrust bombs as indicators of (R and n are the same as in part (a)).
Vulcanian eruption dynamics at Guagua Pichincha (d) Compare your solutions and discuss possible
volcano, Ecuador. Bulletin of Volcanology, 69, reasons for the differences.
281–300, doi:10.1007/s00445–006–0073–6.
Online resources available at www.cambridge.
Yamamoto, H., Watson, I. M., Phillips, J. C. and
org/fagents
Bluth, G. J. S. (2008). Rise dynamics and relative
ash distribution in Vulcanian eruption plumes at • Additional exercises and supporting materials
Santiaguito Volcano, Guatemala, revealed using • Answers to exercises
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
Observations of these, and many other events, winds. Chapter 9 further discusses the differ-
point to the tremendous energy in eruption ences between strong and weak plumes.
columns, which can transport vast quantities Figure 8.2 shows the constituent parts of a
of erupted material, as well as lower atmos- large eruption column. On exit from the vent,
pheric air, high into the atmosphere. To pro- the gas–particle mixture emerges into the gas-
vide constraints on plume models, pertinent thrust region or momentum-driven jet, which typ-
data from historical eruptions include the vol- ically extends to 1–2 km above the vent. The
ume erupted and the duration of the events, mixture is hot and dense, with a temperature in
as well as estimates of the height of rise of the range 800–1200˚C, density up to several tens
the material into the atmosphere. These data times that of the air, and with pressures as large
enable comparison of model predictions of as 10–100 times atmospheric pressure, depend-
column rise height as a function of eruption ing on the geometry of the vent and the mass of
rate with field observations. exsolved gas in the erupting mixture. The mix-
The purpose of this chapter is to review some ture initially undergoes rapid decompression
of the present models and understanding of the to atmospheric pressure, producing significant
dynamics that control the evolution of volcanic expansion. Subsequently, the material continues
eruption columns. We start with a description to ascend into the atmosphere as a highly turbu-
of the dominant physical processes, in order to lent flow, with a large mean speed typically in
provide context for the quantitative models that the range 100–150 m s–1. The mixture of dense
are subsequently discussed. We present some of particles and gas decelerates as it ascends, and
the important predictions from these models some of the larger particles may separate from
and discuss some of the models’ limitations. the ascending mixture and follow ballistic tra-
A brief review of numerical and experimental jectories to the ground, while the bulk of the
models is then given, illustrated with some field flow continues upwards. The column commonly
observations. By way of comparison, we also exhibits lateral gradients in density and momen-
explore some of the controls on the dynamics of tum, with the material at the edges mixing very
phreatomagmatic eruptions (discussed in greater effectively with the surrounding air, which is
detail in Chapter 11). At the low-intensity end of then transferred to the inner core of the column
the explosive continuum, the dynamics of lava over time through shear mixing. As the air is
fountains are described, before we conclude stirred into the ascending mixture by€the turbu-
with some suggestions for future modeling and lence, there is rapid heat transfer from the fine
field work. particles to the air. As a result, the gas content
of the ascending mixture increases and this gas
is heated to temperatures hundreds of degrees
greater than that of the surrounding air. The gas
8.2╇ ╇ Physical processes component of the ascending mixture therefore
becomes much less dense than the surrounding
There are a range of styles of explosive volcanic air, even though in the lower parts of the col-
eruption, ranging from short-lived transient umn, the bulk density of the gas and solid mater-
explosions (see Chapters 6 and 7) to more sus- ial is greater than that of the air.
tained hawaiian lava fountains, and subplinian, If the initial upward speed is sufficiently high
plinian, and ultraplinian eruptions involving the or the eruption rate is sufficiently small, then
ejection of very large volumes of ash and gas. In the continuing entrainment of air may eventu-
this chapter, we primarily consider plinian-style ally lead to the bulk density of a part of the flow
eruptions, saving discussion of hawaiian-style becoming less than that of the surrounding air.
fountains to Section 8.4. Furthermore the At this point, the buoyancy will drive the mix-
description and models that follow apply pri- ture upwards, and the overall upward momen-
marily to strong plumes, which are sufficiently tum of the flow will increase. As this buoyant
vigorous that they are unaffected by atmospheric mixture continues to rise convectively, more
external water, the tendency towards thorough variation with height (Woods, 1988; Sparks et
fragmentation may increase the efficiency of al., 1997). This approach has proved very useful
heat transfer from fine particles to the gas phase, in exploring the dynamics of eruption columns
and hence drive higher columns (Chapter 11). and in assessing the different controls on the
However, when larger quantities of water are flow.
involved, the initial thermal energy of the solid Third, it is possible to formulate a set of
may be depleted through heating and vapori- equations that govern the conservation of mass,
zation of the water, and this in turn may sup- momentum, and energy at each point in the
press the subsequent formation of a buoyant flow. This includes a parameterization of the
column, leading instead to vapor-rich or wet local effects of the turbulence, and the use of
PDCs. We discuss some of these effects later averaged quantities for the bulk properties of
in the chapter. We also describe some recent the multiphase flow at each point. These equa-
numerical modeling that has captured much tions are then solved numerically, using an
of the dynamics of eruption columns, and pro- axisymmetric coordinate system (e.g., Neri and
vides new insights into the phenomena at the Dobran, 1994; Neri and Macedonio, 1996; Neri
transition from plume-forming to flow-forming et al., 1998) or as a three-dimensional flow (e.g.,
behavior. Suzuki et al., 2005), by resolving the relevant
dynamic scales to capture air entrainment and
mixing, and thereby to examine the motion of
the column.
8.3╇ ╇Quantitative modeling of This review focuses primarily on the first
eruption columns two approaches. However, Section 8.3.5 dis-
cusses some of the recent developments in
In order to develop a quantitative description of numerical simulation of eruption columns,
the motion of volcanic eruption columns, which which replicate the predictions of the sim-
are highly turbulent, time-dependent phe- pler integral models (e.g., Woods, 1988), but
nomena, there is a range of possible modeling also provide new insights into the mixing and
approaches. First, one can follow the principles partitioning of the flow between plume- and
of dimensional analysis and determine some PDC-forming regimes. These numerical simu-
of the key integral properties of the column in lations have evolved over the past ten years
terms of the effective buoyancy flux in the col- from axisymmetric steady-state models to the
umn and the stratification of the environment present time-dependent three-dimensional
(e.g., Morton et al., 1956). This provides a simpli- models.
fied but powerful means of assessing some of
the main properties of the eruption, including 8.3.1â•… Dimensional analysis
the eruption rate and/or height of rise of the The motion of turbulent buoyant plumes has
eruption column. been studied in a variety of contexts over the
Second, one can build a time-averaged model past 50 years, following the pioneering work of
of the evolution of the properties of the col- Morton et al. (1956), in which some of the under-
umn, accounting for the effects of turbulence lying physical principles of their motion were
by parameterizing the mixing rate between the first presented. In that work, it was recognized
ambient air and the ascending mixture in the that when a localized source of buoyancy pro-
column. In such models, integral properties of duces a quasi-steady turbulent buoyant plume
the column are related at each height in the col- in a stratified environment, the height of rise
umn. In developing this modeling approach, it of the plume depends on the buoyancy flux, B,
is helpful to examine the horizontally averaged defined as
properties of the column (Morton et al., 1956).
∆ρ
The principles of mass, momentum, and energy B= gV , (8.1)
ρo
conservation are then used to determine their â•…
where V is the volume flux of the source, and This heat flux evolves as the mixture entrains
Δρ is the difference between the density of the the relatively dense air from the lower atmos-
ambient fluid ρo at the level of the source, and phere and carries it with the erupted material
the density of the source fluid. (Notation is sum- upwards into the atmosphere, thereby gener-
marized in Section 8.6.) The stratification is ating potential and kinetic energy. However, in
measured by the Brunt–Väisälä frequency, the lower part of the column, we can define an
effective buoyancy flux associated with this heat
g ∂ρ a
N= − , where ρa is the density of the flux as
ρ o ∂z
Q Hg (8.4)
ambient fluid, and ρo is taken as 1 kg m–3. By B= .
dimensional analysis the height of rise then ρ a cTa
takes the form
We can then use Eq. (8.2) to determine the
H = λ B1/ 4 N −3 / 4 , (8.2) height of rise of the eruption column (see Fig.
8.3). In evaluating the Brunt–Väisälä frequency
where the constant λ is ~5, as deduced from
N, some approximations are required owing
experimental measurements. This model is
to the structure of the atmosphere. The lower
based on an implicit assumption that the source
atmosphere is relatively well mixed and so N
flux is small compared to the flow rate in the
is very small, while the stratosphere is much
column, which is typically the case sufficiently
more stably stratified, and N is larger, which
far from the source, as the ascending mate-
will arrest the motion of the plume. In applying
rial entrains and convects air upwards. It also
the model, using a fixed value for N is a sim-
assumes that the buoyancy flux of the erupt-
plification likely to lead to errors in predicting
ing material is conserved in a non-stratified,
the rise height of both small eruption columns
well-mixed environment, so that mixing of the
that remain within the troposphere, and power-
source fluid with the ambient fluid causes the
ful eruption columns that penetrate well into
buoyancy to decrease at a rate that is inversely
the stratosphere. One of the benefits of using an
proportional to the volume flux.
integral model of the plume properties at each
This latter assumption does not apply to the
height in the plume (see Section 8.3.2) is that
motion of a volcanic eruption column however,
the variation with height of the atmospheric
since the buoyancy is not conserved. Indeed,
stratification can be included in the model of
once the material has emerged from the vent
the turbulent buoyant plume.
and decompressed to atmospheric pressure, the
There are a number of compilations of col-
material is initially dense compared to the air,
umn rise height and eruption rate estimates
but through mixing and heat transfer, it becomes
for various historic eruptions (Sparks et al.,
less dense than the air. However, in a uniform
1997). Some of the key eruptions include Mount
environment the energy flux is conserved, and
Pinatubo, 1991, for which the eruption column
this can be used to adapt Eq. (8.2) for application
rose over 34 km with an eruption rate of 4 × 106
to a volcanic eruption column, as recognized by
m3 s–1; Vesuvius AD 79 (the “grey pumice phase”),
Wilson et al. (1978). For this purpose, it is neces-
for which the column rose 32 km with an esti-
sary to define the effective buoyancy flux at the
mated eruption rate of 6 × 105 m3 s–1; Quizapu,
base of the plume in terms of the heat flux. If, at
Chile, 1932, which generated a 27–30 km high
the volcanic vent, the source mass flux is Q, the
column with an estimated eruption rate of
specific heat is c and the temperature is To, then
6 × 105 m3 s–1; and a number of smaller erup-
the heat flux measured relative to the tempera-
tions, including La Soufriere, St. Vincent, 1902,
ture of the air at the ground Ta is
which produced a column 15.5–17 km high
with an estimated eruption rate of 1.1–1.5 × 104
Q H = Qc (To − Ta ). (8.3) m3 s–1. Figure 8.3 shows a comparison of these
data (symbols) with the dimensional scaling
d
dz
( ρ u2b2 ) = b2 g (ρa − ρ ), (8.6)
Figure 8.4╇╇ Schematic illustration of the variables used in while the steady flow energy equation, which
the theoretical model of an eruption column. The plume of describes the conservation of enthalpy flux in
radius b, density ρ, temperature T, and gas mass fraction n the eruption column, takes the form
rises with velocity u through an atmosphere of density ρa and
temperature Ta. All parameters vary with height, z, but are
d 2 u2
assumed to be constant across the width of the plume at any ρub c (T − To ) + + gz
given height.
dz 2 (8.7)
= 2 ε bu ρa (c a (Ta − To ) + gz ) ,
flows in which there is internal generation of
buoyancy through heating and expansion of where c is the specific heat, and subscripts a and
the entrained air (Caulfield and Woods, 1995), o represent, respectively, properties of the ambi-
the entrainment coefficient can change owing ent air and a reference or initial value (e.g., the
to the evolving structure of the turbulence in ambient temperature at the ground). Equation
the flow. This can be described in the model- (8.7) represents an expression for the rate of
ing through inclusion of a variable entrainment change of the averaged internal energy, kinetic
coefficient (Woods, 1988, Kaminski et al., 2005, energy, and potential energy of the ascending
Suzuki et al., 2005). Changes in the entrain- mixture (terms on left-hand side of Eq. 8.7),
ment rate lead to differences in the quantita- which results from the entrainment of internal
tive predictions of the model, but many of the energy and potential energy associated with
key phenomena can be identified using a con- the ambient fluid (terms on right-hand side).
stant value for the entrainment coefficient. It is Equation (8.7) may be expressed in the form
worth noting that there are many approxima- d
ρ ub2c (T − To ) = 2ε bu ρ a c a (Ta − To )
tions in the model, including the assumptions dz (8.8)
of thermal equilibrium and momentum equili- d
bration between the gas and particles, which − g ρ ub2 − ( ρ u 3 b2 ) .
dz
may evolve with time if the particle size distri-
bution evolves, as well as the possible effects The specific heat of the material in the column
of particle fallout from the umbrella cloud and is given by the mass average of the solid and gas
subsequent re-entrainment. An important role specific heats, denoted by subscripts s and g,
of integral models, is to capture some of the key
processes, and explore how they may impact c = ( (1 − n) c s + nc g ) , (8.9)
â•…
the flow dynamics and hence the overall evo-
lution of the flow. It is also possible to replicate where specific heat of the gases, cg, can be found
specific eruption histories with these models, from the weighted average of the specific heats
in order to predict eruption rates or column of volcanic gas and air. The gas mass fraction in
rise heights. However, inferences from such the column follows from the equation for con-
“simulations” should include error estimates servation of solid material in the column
d
dz
( (1 − n) ρ ub2 ) = 0 (8.10)
1
c g = (n − no )c a + no (1 − n)c vg , (8.11)
n (1 − no ) Figure 8.5╇╇ Model calculations showing the evolution of the
ascent velocity of the plume as a function of height above the
vent. Curves are given for initial speeds of 50, 75, and 200 m
where ca and cvg are the specific heats of the air s-1. The lowest speed does not allow the mixture to become
and volcanic gas respectively, and no is the ini- buoyant before collapse of the plume occurs.
tial gas mass fraction in the erupting material.
Finally, the bulk density of the mixture of gas
and particles is given by far above the vent and the column collapses to
produce a dense fountain that sheds pyroclastic
−1
n 1− n density currents that spread along the ground.
ρ= + , (8.12) At the larger eruption speed of 75 m s–1, the
ρ g ρ s
material again decelerates under gravity in the
region just above the source, but then becomes
where the gas density ρg is given by the mass buoyant as a result of entrainment of air. At this
average of the air and volcanic gases. point, the mixture begins to accelerate upwards
The above system of equations can be solved since it has such a low speed. However, as it
numerically to determine the motion through entrains progressively more air, the momen-
an atmosphere with given vertical temperature, tum associated with the entrainment of the
pressure, and density profiles. In many simu- air begins to dominate the upward buoyancy
lations, including the calculations presented force, and the velocity decreases towards the
here, a simple model atmosphere is used, such maximum rise height. At the largest initial vel-
as that described by Gill (1981). Solutions of the ocity of 200 m s–1, the flow decelerates rapidly
equations corresponding to a dry eruption col- above the source, but then the mixture becomes
umn composed of fine ash, air and, volcanic gas, buoyant and continues to ascend high into the
identify some of the dominant processes in oper- atmosphere.
ation in a volcanic plume. Although the model The critical velocity for column collapse can
is simplified, especially with (i) the entrain- be calculated from the model. We note that there
ment and mixing processes in the lower part of is some variation with such predictions depend-
the column being approximated with a single ing on the precise model for entrainment in the
coefficient, and (ii) the fact that in this region, lower part of the column. Smaller values of the
the flow may divide into a dense PDC-forming entrainment coefficient lead to prediction of
component and a buoyant eruption column, collapse in more cases, owing to the reduction
rather than remaining as a coherent flow, the of mixing and hence inhibition of buoyancy gen-
model is able to delineate broadly the different eration. Furthermore, the assumption that the
styles of behavior observed in nature. Figure 8.5 flow remains well mixed and unidirectional in
illustrates the variation of plume velocity with this lower part of the eruption column is some-
height for three different initial speeds of 50, what simplified; in practice the flow may become
75, and 200 m s–1. At the lowest initial velocity, unsteady as conditions approach those for col-
the erupting material falls to zero velocity not lapse of the column. Therefore, calculations of
the flow. A simplified version of the integral and this depends on the thermal diffusivity κ
model presented above, including a constant and the particle diameter, d. Typically, ther-
entrainment coefficient, was shown to agree mal equilibration occurs within a time of order
with these experiments (Woods and Caulfield, d2/κ. In order to understand the significance
1992). However, in the volcanic problem, which of this timescale compared to the timescale of
includes heat transfer and mixing, the picture is the plume, we compare the conduction time
more complex. Woods and Caulfield (1992) also scale with the time scale of ascent in the col-
demonstrated that below a critical mass flux, umn, which is of order 1/N. For the atmosphere,
the MEG is unable to mix with sufficient water 1/N has a value of 100–200 s, and pumice clasts
to become buoyant, and instead it remains have a thermal diffusivity of order 10–7 m2 s–1,
dense and forms a collapsing fountain. Near the so that fragments smaller than 1 mm equili-
conditions for collapse, the experimental flows brate in less than one-tenth of the ascent time
became highly unsteady, with intermittent peri- of the column. This implies that particles larger
ods of plume- and flow-forming behavior; the than about 1 mm contribute progressively less
exterior of the ascending mixture commonly of their thermal energy during ascent in the
entrained and mixed efficiently with the sur- plume. In principle, 3 mm fragments are the
rounding fluid, while the interior of the flow largest that are able to fully equilibrate prior to
remained denser and less well-mixed, leading to reaching the plume top. In more coarse-grained
these instabilities. As the flow collapsed asym- eruptions, this lack of thermal equilibrium may
metrically, the continuing flow became buoyant have a profound impact on the plume ascent
and lifted off the horizontal surface in a manner and also on the conditions for column collapse.
analogous to PDCs. Section 8.4 discusses some of the dynamics of
lava fountains, and shows how the thermal and
8.3.3â•… Non-equilibrium effects dynamical disequilibria of such coarse pyroclast
The model described in Section 8.3.2 is based populations lead to behavior very different from
on the assumption of equilibrium, in terms convecting columns.
of both€ the dynamical and thermal coupling In order to account for thermal disequi-
between the particles and the gas. For eruptions librium effects, the ascent of the particles,
in which the erupting material does not frag- including the vertical dispersion owing to the
ment as efficiently, there will be some disequi- turbulence, needs to be modeled along with the
librium between the coarser-grained fragments time-dependent heat transfer to the surround-
and the gas. Thermal disequilibrium occurs because ing material. This involves a non-trivial devel-
heat is transferred less efficiently between coarse opment of the model to quantify the vertical
particles and the gas. This reduces the supply dispersion within a turbulent plume and also
of heat and hence may suppress the generation to keep an inventory of the different particles
of buoyancy as the plume ascends. In turn, this in the plume in terms of their size and time of
may lead to lower column heights. Dynamical dis- release, and hence their temperature. Although
equilibrium occurs as the larger particles, under this is possible, in the context of the simplified
the influence of gravity, decouple aerodynamic- integral models of an eruption column, a sim-
ally from the gas. This reduces the load of dense pler heuristic approach has been developed to
particles and so may assist in the buoyancy gen- explore the impact of disequilibrium through
eration low in the column. However, higher in introduction of a parameter, f, which reduces the
the column, the lack of solid material reduces fraction of the particles that are in good thermal
the overall heat flux, and hence may lead to contact with the gas. This approach essentially
lower column heights, even though the mass of leads to a bimodal distribution of particles from
solid material is reduced. a thermal perspective, with a fraction in equi-
Thermal equilibration between the solid librium and the remainder in disequilibrium
particles and the air is ultimately controlled by throughout their ascent in the column; this cor-
the diffusion of heat within the solid fragments, responds to particles smaller than ~ 0.5–1.0 mm
ρu
φ (top) = φ (source) exp t ( rme 2 − ro2 ) . (8.17)
Q
8.3.6â•… Co-ignimbrite clouds speeds upon exit from the vent. The clasts read-
One of the other fascinating phenomena that ily decouple from the low-velocity gas phase,
have been observed to occur in large explosive so instead of erupting as a high-speed, effect-
eruptions is the formation of large convecting ively single-phase flow as for the more explo-
plumes above pyroclastic density currents that sive counterparts, the erupting material forms
are generated by collapsing fountains. These a multiphase flow. This can lead to a range of
plumes arise through the entrainment of air eruption phenomena from ballistic trajectories
and sedimentation of particles from the ground- for the coarser fragments, forming incandescent
hugging flow, which results in the continuing lava fountains (when erupted from point source
flow becoming less dense than the air and lift- vents) or curtains of fire (when erupted from lin-
ing off to form buoyant plumes. Other plumes ear fissures), to buoyant plume-forming behav-
arise from the elutriation of the finer size frac- ior for the finest particle fraction (Fig. 8.1(b)).
tion as the current continues to flow. The result- Fountaining commonly gives way to passive
ant dynamics of such co-ignimbrite clouds is lava effusion of less volatile-rich magma as the
analogous to the convective region of eruption eruption progresses.
columns, but the material may lift off far from A range of landforms and deposits can be pro-
the vent (Dobran et al., 1993). If the material in duced from lava fountaining activity, depending
these flows cools sufficiently as it spreads along on the local temperature of clasts on landing and
the ground, for example as a result of mixing the accumulation rate of the clasts (Head and
with water, then the continuing flow may be Wilson, 1989). Temperature and accumulation
unable to generate buoyancy (Koyaguchi and rate in turn depend on the clast trajectories (a
Woods, 1993). function of fountain size and structure), which
are ultimately related to the magma mass flux
and volatile content (Fig. 8.14(a,b)). For example,
relatively large volatile contents lead to high
8.4╇ ╇Basaltic systems and eruption velocities, tall fountains, and wide-
lava-fountaining eruptions spread dispersal from the vent. Particles land-
ing far from the vent will be relatively cool on
Basaltic magma is less viscous and has lower landing, whereas coarse clasts landing close to
volatile contents than the intermediate to felsic the vent will be hot and fluid. Therefore a range
magmas that tend to produce convecting erup- of deposits may be observed, from rheomorphic
tion columns. Because of the low viscosity, the lava flows produced by coalescence of rapidly
exsolving gas phases can decouple more readily accumulating, hot clasts in proximal locations,
from the melt as it ascends to the surface, so that to cone-forming agglutinated spatter and/or brit-
for low magma ascent speeds, intermittent emis- tle scoria at increasing distances, and sheet-like
sion of large gas slugs can produce strombolian ash fall deposits dispersed distally from the fine-
explosions (Chapter 6). At greater magma rise grained component rising above the fountain
rates, the bubbles and magma remain dynam- (Fig. 8.14(c)). As an illustration, during the 1995
ically coupled, and the low viscosity prevents eruption of Cerro Negro (Hill et al., 1998), an
significant bubble overpressures from building eruption column reached an altitude of about 5
up as the bubbly magma decompresses during km and transported small particles with median
ascent (Chapter 4, Section 4.3.2). The lack of sig- size of 0.7 mm to > 30 km downwind of the vol-
nificant bubble overpressure, together with the cano, while also contributing to the growth of
low total volatile content, means that fragmen- a near-source cone from larger volcanic blocks
tation, if it occurs, is less vigorous than for more and bombs up to 2 m in diameter.
silicic systems. As a result, a coarser pyroclast Lava fountains range from a few meters
size distribution is produced (with clasts typic- high to several hundred meters in height. For
ally ranging from millimeters to tens of centim- example, the 1959 eruption of Kilauea Iki in
eters), and gas expansion produces relatively low Hawaii (Fig. 8.1(b)) produced fountains close to
600 m high (Richter et al., 1970). Fountain height, eruptions, and Woods (1993) extended this to
Hf, can be related to the vent speed, uo, given in account for the effects of the entrainment of
terms of a simple ballistic relation, neglecting moisture. Such calculations show that even if
air resistance for sufficiently large particles, only a small fraction of the mass erupting in
a basaltic fissure eruption leads to formation
uo = 2 gH f , (8.18) of an ash plume, the plume may rise several
kilometers through the lower atmosphere
which implies that pyroclasts are being ejected because the air is only weakly stratified. The
from the vent with speeds up to 110 m s–1. potential for such plumes to inject large quan-
Fountain heights and eruption velocities can in tities of aerosols high into the atmosphere has
turn be linked to the exsolved magma volatile been proposed as one process whereby mas-
content driving the eruption (Head and Wilson, sive flood basalts can impact climate. Since
1987). The coarse size of most clasts means that, such plumes typically only carry a small flux of
upon eruption, they are slipping relative to the erupting material, in humid environments, the
gas phase, which suggests that gas eruption entrainment and vertical transport of water
speeds are significantly greater than the veloci- may be a key factor for increasing the height of
ties of magma fragments calculated from foun- rise of such plumes.
tain heights (Wilson, 1999). In considering sustained explosive activity, it
While coarse clasts fall rapidly and thus is important to note that there is a range of phe-
are unable to contribute thermally to driv- nomena, from eruption of very coarse-grained,
ing a buoyant column, the gas phase and fine fluid lava fountains to the highly fragmented
particles may rise above the lava fountains to mixture that forms ash plumes. This range may
form a small ash plume. Stothers et al. (1986) be considered as a continuum of activity, depend-
estimated the height of rise of these plumes in ent on the degree of fragmentation (and hence
terms of the mass flux supplied from basaltic efficiency of heat transfer between particle and
Δρ density difference between ambient fluid Hill, B. E., Connor, C. B., Jarzemba, M. S. et al. (1998).
at source level and the source fluid 1995 eruptions of Cerro Negro volcano, Nicaragua,
(kg m–3) and risk assessment for future eruptions. Geological
ρa density of ambient air (kg m–3) Society of America Bulletin, 110, 1231–1241.
ρ g density of gas phase (kg m–3) Kaminski, E. and Jaupart, C. (2001). Marginal stability
of atmospheric eruption columns and pyroclastic
ρo density of ambient fluid at source level
flow generation. Journal of Geophysical Research,
(kg m–3)
106(B10), 21785–21798.
ρs particle density (kg m–3)
Kaminski, E., Tait, S. and Carazzo, G. (2005).
Turbulent entrainment in jets with arbitrary
buoyancy. Journal of Fluid Mechanics, 526, 361–376.
References Koyaguchi, T. and Woods, A. W. (1996). On the
explosive interaction of magma and water, Journal
Bursik, M. I. and Woods, A. W. (1991). Buoyant, of Geophysical Research, 101, 5561–5574.
superbuoyant and collapsing eruption columns. Morton, B., Taylor, G. I. and Turner, J. (1956).
Journal of Volcanology and Geothermal Research, 5, Turbulent gravitational convection from
347–350. maintained and instantaneous sources. Proceedings
Caulfield, C. C. P. and Woods, A. W. (1995). Plumes of the Royal Society of London, A231, 1–23.
with nonmonotonic mixing behavior. Geophysical Neri, A. and Dobran, F. (1994). Influence of eruption
and Astrophysical Fluid Dynamics, 79, 173–199. parameters on the thermofluid dynamics of
Dobran, F., Neri, A. and Macedonio, G. (1993). collapsing volcanic columns. Journal of Geophysical
Numerical simulation of collapsing volcanic Research, 99, 11,833–11,857.
columns. Journal of Geophysical Research, 98(B3) Neri, A. and Macedonio, G. (1996). Numerical
4231–4259. simulation of collapsing volcanic columns with
Ernst, G., Sparks, R. S. J., Carey, S. N. and Bursik, particles of two sizes. Journal of Geophysical Research,
M. I. (1996). Sedimentation from turbulent jets 101(B4), 8153–8174.
and plumes. Journal of Geophysical Research, 101, Neri, A., Papale, P. and Macedonio, G. (1998). The
5575–5589. role of magma composition and water content
Gill, A. (1981). Atmosphere−Ocean Dynamics. in explosive eruptions: II. Pyroclastic dispersion
International Geophysics Series. New York: dynamics. Journal of Volcanology and Geothermal
Academic Press. Research, 87, 95–115.
Glaze, L. S. and Baloga, S. M. (1996). Sensitivity Richter, D. H., Eaton, J.P., Murata, K. J., Ault, W.
of buoyant plume rise heights to ambient U. and Krivoy, H.L. (1970). Chronological Narrative
atmospheric conditions: Implications for volcanic of the 1959–60 Eruption of Kilauea Volcano, Hawaii.
eruption columns. Journal of Geophysical Research, United States Geological Survey Professional
101(D1), 1529–1540. Paper 537-E.
Glaze, L. S., Baloga, S. M. and Wilson, L. (1997). Sparks, R. S. J. (1986). The dimensions and dynamics
Transport of atmospheric water vapor by volcanic of eruption columns. Bulletin of Volcanology, 48,
eruption columns. Journal of Geophysical Research, 3–15.
102(D5), 6099–6108. Sparks, R. S. J., Bursik, M., Carey, S. et al. (1997).
Head, J. W. and Wilson, L. (1987). Lava fountain Volcanic Plumes. Chichester: John Wiley.
heights at Pu’u ‘O’o, Kilauea, Hawaii: Indicators of Stothers, R., Wolff, J. A., Self, S. and Rampino, M.
amount and variations of exsolved magma volatiles. (1986). Basaltic fissure eruptions, plume heights
Journal of Geophysical Research, 92, 13,715–13,719. and atmospheric aerosols. Geophysical Research
Head, J. and Wilson, L. (1989). Basaltic pyroclastic Letters, 13, 725–728.
eruptions: Influence of gas-release patterns Suzuki, Y., Koyaguchi, T., Ogawa, M. and Hachisu, I.
and volume fluxes on fountain structure, and (2005). A numerical study of turbulent mixing in
the formation of cinder cones, spatter cones, eruption clouds using a three dimensional fluid
rootless flows, lava ponds and lava flows. Journal of dynamics model. Journal of Geophysical Research,
Volcanology and Geothermal Research, 37, 261–271. 110(B08), 201–219.
Valentine, G. and Wohletz, K. (1989). Numerical Woods, A. W. and Bursik, M. I. (1991). Particle fallout,
models of Plinian eruption columns and thermal disequilibrium and volcanic plumes.
pyroclastic flows, Journal of Geophysical Research, Bulletin of Volcanology, 53, 559–570.
94(B2), 1867–1887. Woods, A. W. and Caulfield, C. P. (1992). A laboratory
Veitch, G. and Woods, A. W. (2002). Particle recycling study of explosive volcanic eruptions. Journal of
in volcanic eruption columns. Bulletin of Volcanology, Geophysical Research, 97, 6699–6712.
64, 31–39.
Walker, G. P. L., Self, S. and Wilson, L. (1984).
Tarawera 1886, New Zealand€– a basaltic plinian Exercises
fissure eruption. Journal of Volcanology and
Geothermal Research, 21, 61–78. 8.1 Estimate the height of rise of a plinian eruption
Williams, S. N. (1983). Plinian airfall deposits of column rising through an environment with a
basaltic composition. Geology, 11, 211–214. Brunt–Väisälä frequency of 0.01 s–1 if the eruption
Wilson, L. (1976). Explosive volcanic eruptions€– III. rate Q is 106 kg s–1. Use typical values for atmos-
Plinian eruption columns. Geophysical Journal of the pheric temperature and density at sea level, and
Royal Astronomical Society, 45, 543–556. assume a rhyolite eruption temperature. How
Wilson, L. (1999). Explosive volcanic eruptions − X. will this change if the particle size gradually
The influence of pyroclast size distributions and increases with time through the eruption?
released magma gas content on the eruption 8.2 As basaltic fissure eruptions evolve, there is
velocities of pyroclasts and gas in Hawaiian and often some flow localization, leading to discrete
Plinian eruptions. Geophysical Journal International, axisymmetric fountains rising from one or more
136, 609–619. point source vents. If 10% of the material from
Wilson, L. Sparks, R. S. J., Huang, T. C. and a basaltic eruption rises to form an ash plume
Watkins, N. D. (1978). The control of volcanic above each lava fountain, and the plume has a
column heights by eruption energetics and height 2.5 km, estimate the eruption rate for
dynamics. Journal of Geophysical Research, 83(B4), each fountain. Assume N = 0.01 s–1 and similar
1829–1836. atmospheric characteristics to Exercise 8.1.
Woods, A. W. (1988). The fluid dynamics and
thermodynamics of eruption columns. Bulletin of Online resources available at www.cambridge.
Volcanology, 50, 169–193. org/fagents
Woods, A. W. (1993). A model of the plumes above • Answers to exercises
basaltic fissure eruptions. Geophysical Research • Additional reading
Letters, GL01215, 1115–1118. • Additional figures
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
The preservation of tephra deposits is typ- or from an extended source (i.e., fissure eruption
ically incomplete. In most eruptions proximal or pyroclastic density current), eventually devel-
areas are buried or collapsed, and distal depos- ops into a turbulent buoyant current whose
ition occurs in the sea or becomes eroded. For dynamics are strongly controlled by the degree
small eruptions, the whole tephra deposit may of interaction with the atmosphere. If the plume
be eroded away within a few years of deposition. upward velocity is much stronger than the wind
As a result, empirical, analytical, and numerical velocity, the initial jet phase (gas thrust) evolves
models have been developed to allow quantita- into a vertical buoyant column that then even-
tive interpretation of tephra deposits and to fully tually spreads laterally as a gravity current (i.e.,
understand the nature of ancient eruptions that umbrella cloud) around the neutral buoyancy level
are incompletely preserved. Dedicated analytical Hb (i.e., strong plume; Fig. 9.1(a)). In contrast, if the
and numerical models have also been produced wind velocity is much stronger than the plume
to investigate plume dynamics and particle upward velocity, the turbulent current will be
sedimentation, and to provide long-term assess- bent over above the basal jet before spreading
ments for land-use planning and rapid response laterally around Hb (i.e., weak plume; Fig. 9.1(c)).
during volcanic crises. Model validations have It is important to distinguish between vigorous
shown good agreement with field data, which and low-energy weak plumes: both plumes are
justifies the use of these models for hazard bent over by the wind but they are characterized
applications. by different energetics (e.g., steepness of plume
This chapter describes (1) empirical and trajectory relative to wind speed). Typically, vig-
analytical models used to determine eruption orous weak plumes characterize the beginning of
parameters, such as column height, eruption low-intensity sustained eruptions (e.g., Ruapehu,
duration, magnitude, and intensity, (2) analyt- 17 June 1996; Bonadonna et al., 2005a), whereas
ical and numerical models developed for the low-energy weak plumes characterize the last phase
study of the dynamics of volcanic plumes and of an eruption when wind eventually dominates
particle sedimentation, and (3) models com- and the cloud starts propagating as a lens of aero-
monly used for hazard assessments and fore- sol (e.g., Mount St. Helens, 22 July 1980; Sparks
cast of plume spreading. Model assumptions et al., 1997).
and caveats are discussed, and a key case study Volcanic clasts (juveniles and lithics) are car-
is presented to facilitate comprehension of the ried up within the turbulent current according
application of the models described (the 22 July to their settling velocity, which depends on both
1998 explosive eruption of Mt Etna, Italy; Cole particle and atmospheric characteristics. When
et al., 1998; Scollo et al., 2008a). particle settling velocities are larger than the
upwards component of the turbulent current,
they fall out and are advected by local winds.
Particles that are sufficiently small will typically
9.2╇ ╇Plume dynamics and particle aggregate into micron- to millimeter-sized clus-
sedimentation ters having greater settling velocities (Sparks
et al., 1997). In addition, the deposition of fine
Before describing the models used to character- particles is also enhanced by pronounced con-
ize tephra dispersal and deposits, it is important vective instabilities and mammatus that often
to understand some basic concepts of plume form at the base of the sedimenting turbulent
dynamics and particle sedimentation (see current (Bonadonna et al., 2002b; Durant et al.,
Chapter 8 for a detailed review of plume dynam- 2009). As a result, the characteristics of tephra
ics). Volcanic plumes are typically associated deposits are the result of plume dynamics (e.g.,
with explosive activity and consist of a mixture plume height, velocity profile, weak-plume vor-
of lithics (wall rock), volcanic gas, and juvenile ticity), particle parameters (e.g., size, density and
particles (fragmented magma), which, whether shape), atmospheric characteristics (e.g., wind
generated from a point source (i.e., single vent) field, atmospheric density and viscosity) and
WIND ADVECTION
Hb
Tu
the eruption (Carey and Sparks, 1986). Isopach/ a result, the logarithm of tephra thickness can
isomass and isopleth maps can also be used to be described by straight lines (i.e., exponential
determine vent location and to classify eruptive segments) when plotted against distance from
style (Walker, 1973; Walker, 1980; Pyle, 1989). vent or square root of the area enclosed by each
Mass eruption rate and the duration of the sus- isopach ( A ):
tained phase of the eruption can be calculated
from plume height and erupted mass, respect- T = To exp (−k A ) (9.1)
ively (Sparks, 1986; Wilson and Walker, 1987;
Carey and Sigurdsson, 1989). Inferences of frag- where To is the maximum thickness of the
mentation mechanisms can also be made from deposit and k defines the rate of thinning of the
the study of particle sizes (Kaminski and Jaupart, deposit (i.e., slope of the associated exponential
1998; Neri et al., 1998; Zimanowski et al., 2003). segment). All notation is summarized in Section
The empirical and analytical models used for 9.8. Assuming that isopachs have elliptical
these purposes, together with their assumptions shapes, the volume of tephra deposit is:
and limitations, require thorough analysis to
assess the variability of resulting eruption param- V = 13.08 To bt2
(9.2)
eters. This is crucial not only because these erup-
tion parameters are used to characterize volcanic where bt = ln(2) (k π ) .
eruptions, but also because they are used as input Fierstein and Nathenson (1992), Pyle (1995),
to numerical models and to construct potential and Bonadonna and Houghton (2005) developed
activity scenarios for hazard assessment. this method to account for abrupt changes in
the rate of thinning of some tephra deposits:
9.3.1â•…Determination of erupted volume
based on the assumption of 2T10 k S + 1 k S + 1
V= + 2T10 2 1 2 − 1 1 2 exp (−k1S1 )
exponential thinning of tephra k12 k2 k1
deposits k S + 1 k S + 1
This approach was introduced by Pyle (1989), + 2T20 3 2 2 − 2 2 2 exp (−k2S2 ) + ... + 2T( n −1)
k3 k2 0
where Tn0, kn, and Sn are the intercept, slope, and field data. In particular, the power-law best fit
position of the break in slope of line segment n. can be described as:
Their approach to estimating volume by defin-
m
ing several exponential segments (i.e., different A
( A)
−m
decay to zero at ground level, and a value above Eruption duration can be determined by divid-
the tropopause of 3/4 of the maximum velocity. ing the total erupted mass by M↜̇ .
Mass eruption rate can also be derived from and 1000 °C, and for both tropical and temper-
the column height H (m) using the analytical ate atmospheres. The model shows good agree-
model of Sparks (1986), which was based on ment with observed data. However, in order to
buoyant plume theory (BPT) (Morton et al., 1956; compile diagrams from more elaborate theory,
Settle, 1978; Wilson et al., 1978) and improved Sparks (1986) made assumptions about tropo-
by accounting for a varying adiabatic lapse rate pause height, surface temperature, tempera-
and atmospheric temperature. As a result, Ṁ and ture gradient, wind profile, and air-entrainment
H show a nonlinear correlation, which strongly models that need to be carefully verified prior
depends on eruption temperature. to application.
Caveats
Equation (9.8) holds only for circular-vent
plumes < 35 km high and is supported by the- 9.4╇ ╇Models based on the
oretical investigations based on BPT (Morton
Advection–Diffusion–
et al., 1956; Wilson and Walker, 1987) which
show that maximum plume height is roughly Sedimentation (ADS) equation
proportional to the fourth root of the heat injec-
tion rate, and therefore to the fourth root of the Models for tephra dispersal are based on the
mass eruption rate (see also Chapter 8, Section mass conservation equation with different
8.3.1). However, there are several limitations in degrees of simplicity, following either Eulerian
extending BPT to eruption plumes in a strati- or Lagrangian formulations. The Eulerian
fied atmosphere where buoyancy flux varies approach describes changes in the fluid at
with height and crosswinds significantly affect fixed points, whereas the Lagrangian approach
plume entrainment (Bursik, 2001; Ishimine, describes changes by following a fluid par-
2006; Carazzo et al., 2008). In addition, Eq. cel along its trajectory. Each approach is use-
(9.8) is strictly valid for a plume temperature ful for different applications. For example,
of ~800 °C, appropriate for andesitic magma. weather forecasting is based on the Eulerian
Basaltic magmas are typically hotter by at least approach (fixed measurement system) because
200 °C, and therefore, to achieve the same col- it uses data from fixed stations around the
umn height the corresponding mass discharge world. The Lagrangian approach is more use-
rates are lower for basaltic magmas (Carey and ful when describing the evolution of a given
Sparks, 1986; Sparks, 1986; Woods, 1988). For material as it moves within a certain fluid (e.g.,
example, Wehrmann et al. (2006) found that chemical modeling). Tephra dispersal is often
C = 295 m kg-1/4 s1/4 in Eq. (9.8) describes the described using both approaches. In particular,
relationship between Ṁ and H for a basaltic models commonly defined as Lagrangian are
plinian eruption of Masaya volcano, Nicaragua based on an Eulerian–Lagrangian approach,
(i.e., Fontana Lapilli). Scollo et al. (2007) and which describes the dynamics of single parti-
Andronico et al. (2008) found C = 247 m kg-1/4 s1/4 cles within an Eulerian flow field. In contrast,
and C = 244 m kg-1/4 s1/4 for the 2001 and 2002 Eulerian models consider the particle phase
eruptions of Etna volcano, respectively. Finally, and the flow field as two continua. The govern-
it is important to bear in mind that Eq. (9.8) is ing equation derived from the mass conserva-
strictly valid only for values of maximum col- tion condition has the following form (Costa
umn height and therefore gives maximum et al., 2006):
values of Ṁ. As a result, the corresponding erup-
tion durations could be underestimated. In con- ∂ cj ∂ ux c j ∂ uy cj ∂ uz c j ∂ v j cj
+ + + − =
trast, the model of Sparks (1986) can easily be ∂t ∂x ∂y ∂z ∂z
(9.9)
applied by using dedicated diagrams compiled ∂u xc ∂u yc ∂u zc
j j j
for sustained buoyant plumes with heights up − − − + S,
∂x ∂y ∂z
to 35 km, eruption temperatures between 400
from this surface (Costa et al., 2006). In a recent the technique used, on deposit exposure, and
study (Folch et al., 2008), FALL3D was general- on data distribution and density. Most erupted
ized to the mesoscale-synoptic domain and cou- volumes derived from field data should be con-
pled with the Weather Research and Forecasting sidered minimum values unless the data sets
(WRF) meteorological model (Michalakes et al., extend hundreds of kilometers from the vent
2005; www.wrf-model.org). (the higher the plume, the larger the deposit
Finally, it is worth mentioning another cat- to be investigated). A review of several meth-
egory of model designed to describe in detail the ods can be found in Froggatt (1982), whereas
evolution of the eruption plume from its rise to the most recent techniques are summarized in
its collapse. Given the high degree of complexity Section 9.3. In addition, recent applications of
of such models, they are usually applied over rela- inversion techniques to analytical models have
tively small horizontal domains (up to few tens of shown promising results. Specifically, mass
kilometers). Some of these models are focused on per unit area and particle size data from indi-
describing the dynamics and thermodynamics of vidual outcrops are inverted through the use
the mixture of hot gases and particles (Dobran of two-dimensional analytical models to derive
et al., 1993; Esposti Ongaro et al., 2007), whereas eruption parameters, such as erupted mass
other models such as ATHAM (Active Tracer High and column height (Connor and Connor, 2006;
Resolution Atmospheric Model; Herzog et╛╛al., Scollo et al., 2008a; Volentik et al., 2010).
1998; Oberhuber et al., 1998) are more appropri- Column height is very important for defining
ate for dealing with the chemical interactions the source term and is related to eruption inten-
and microphysical processes of volcanic and sity (i.e., mass flux). The best evaluation of col-
cloud water, cloud ice, rain, and graupel. umn height comes from well-documented and
calibrated direct observations. Column height
can also be estimated through analysis of satel-
9.5╇ ╇Limitations of input lite images, based on geometry (cloud shadow
clinometry), thermal infrared (IR) data (using
parameters and a cloud-top temperature/temperature profile
parameterizations adopted method), and correlation of cloud trajectory
by ADS models with meteorological motion (cloud stereoscopy)
(Holasek and Self, 1995; Prata and Turner, 1997;
Regardless of the complexity of different sedi- Glaze et al., 1999; Prata and Grant, 2001). Field
mentation models, the reliability and uncertain- studies have also shown that the derivation of
ties of the associated outputs strongly depend on plume height using the method of Carey and
the reliability and uncertainties of input param- Sparks (1986) gives fairly consistent results
eters (i.e., erupted mass, column height, total even for poorly exposed deposits (Wehrmann
grainsize distribution, meteorological data) and et al., 2006). Inversion of two-dimensional analyt-
of the parameterizations used to describe crit- ical models using particle size data also gives a
ical sedimentation processes, such as particle good constraint on plume height (Volentik et al.,
aggregation, particle terminal velocity, and col- 2010). However, it is important to bear in mind
umn dynamics. that even a small uncertainty in plume height
results in an uncertainty about four times larger
9.5.1â•…Input parameters: erupted mass, in mass flux Ṁ, because of the fourth power rela-
grainsize distribution, plume height, tionship between Ṁ and H (Eq. 9.8).
wind profile All tephra dispersal models are strongly
Erupted mass is the most important input dependent on the choice of initial grainsize
parameter and is one of the most difficult to distribution. Nonetheless, even though sev-
derive accurately from field data. In fact, vol- eral methods have been proposed, including
ume (and mass) estimation strongly depends on simple data averages, sectorization, Voronoi
Several studies have shown that particle- where L is the longest particle dimension,
settling velocities strongly depend on particle I is the longest dimension perpendicular to L,
shape (Wilson and Huang, 1979), although for and S is the dimension perpendicular to both L
simplicity particles are typically assumed to be and I. In addition, Riley et al. (2003) considered
spheres, for which terminal velocities can be the determination of particle sphericity from
determined using simple expressions (Kunii two-dimensional images, which permits analy-
and Levenspiel, 1969; Arastoopour et al., 1982). sis of small particles. However, these methods
The settling velocity vj of particles of size dj is are all approximations that need to be tested
obtained from the balance between gravity and thoroughly for application to calculating ter-
air drag. The drag coefficient, Cd, is a function of minal velocities of volcanic particles. As a
the particle shape and the Reynolds number, Re = result, the effect of particle shape on terminal
dj ρa vj /ηa, where ηa is the air dynamic viscos- velocity is a critical factor that remains to be
ity (Pa s). The assumption of spherical particles adequately described. However, the drag coef-
is valid as a first-order approximation only; for ficient strongly depends on particle shape only
non-spherical particles the determination of Cd for relatively large particles (Fig. 9.3). For exam-
is more complicated. Walker et al. (1971) showed ple, the model of Ganser (1993) can be used to
that pumice clasts >5 mm are better described investigate the terminal velocities of irregular
by cylinders than spheres, and Wilson and particles of known shape and diameter ranging
Huang (1979) found that, for particle diameters between 1.5 and 7 cm, and with shape factor F
from 30 and >500 μm, glass and feldspar frag- ( = (I + S) / 2L; Wilson and Huang, 1979) ranging
ments have a very high proportion of flattened between 0.3 and 0.9. Figure 9.3 shows the associ-
particles, whereas pumice clasts have a greater ated drag coefficients for Reynolds numbers Re
variety of shapes, including equant particles. between 0.001 and 106. Note that the shape fac-
Following a review of available methods for tor F does not uniquely constrain elongated and
estimating the drag coefficient of non-spheri- platy particles (i.e., F < 0.7), and that rounded
cal particles, Chhabra et al. (1999) showed that particles can have similar drag coefficients to
the best approach appears to be that of Ganser elongated particles. Furthermore, Figure 9.3
(1993), which uses the equal volume sphere shows that the drag coefficient varies signifi-
diameter and the sphericity ψ of particles, with cantly only for particles falling in the interme-
a resulting overall error within ~16% for Re ran- diate and turbulent regimes (Re > 1). As a result,
ging from 10−4 to 5 × 105. Unfortunately, expres- future studies of the effects of particle shape on
sions for terminal velocity that account for terminal velocity should focus on medium- and
the complexity of irregular particles are com- particularly high-Re particles, i.e., particles with
monly based on particle parameters that are diameters > 63 μm and > 2 mm respectively, for
impractical to measure. For example, particle heights > 10 km above sea level.
surface area necessary to calculate ψ, cannot An accurate description of plume dynamics
be easily determined since this would imply a is crucial for both analytical and numerical mod-
complicated integration over surface elements els (Scollo et al., 2008b). Model results are very
of an irregular particle. For this reason, Wadell sensitive to the choice of velocity profile within
(1933) and Aschenbrenner (1956) introduced the plume, which ultimately controls both the
the concepts of “operational” and “working mass and the grainsize distribution within the
sphericity,” based on the determination of the eruptive column (e.g., Carey and Sparks, 1986;
volume and of the three dimensions of a par- Bursik et al., 1992a). As a first approximation,
ticle respectively: models based on BPT consider a Gaussian dis-
tribution profile across the plume and a monot-
S
2 S I S
2
I
2 onic trend with height (Carey and Sparks,
ψ work = 12.8 3 1 + 1 + + 6 1 + 2 1 + L2 1986; Bursik et al., 1992a; Sparks et al., 1992;
IL I L I
Bonadonna and Phillips, 2003; Bonadonna et al.,
(9.15) 2005a). However, numerical models show more
complicated velocity profiles, and simulations leading to tephra dispersal to the southeast. The
from three-dimensional models of weak plumes highest seismic tremor recorded for this plume-
show no Gaussian cross section at any time or forming lava fountain lasted for about 25 minutes
even as a time-averaged property. The advan- (Aloisi et al., 2002; Coltelli et al., 2006). The asso-
tage of simple column models such as those ciated tephra blanket was sampled between ~3
used in the analytical models described above and 30 km from the vent soon after deposition.
or the steady-state models based on BPT (Woods, As a result, a detailed isomass map was compiled
1988; Bursik, 2001; Ishimine, 2006) is their com- (Fig.€ 9.4), and a maximum column height of
putational speed and flexibility compared to 11 km a.s.l. and a maximum wind speed at the
the complex three-dimensional time-dependent tropopause of 10–30 m s−1 were determined
descriptions of plume dynamics (Dobran et al., using the method of Carey and Sparks (1986)
1993; Herzog et al., 1998; Oberhuber et al., 1998; (Andronico et€al., 1999; Table 9.1). The subplinian
Esposti Ongaro et al., 2007). As a result, for some character of this eruption is suggested by com-
regimes and applications, a challenge remain- parison with other tephra deposits on a semi-
ing for the volcanology community is to develop log plot of T vs. A (Fig. 9.5). Plinian eruptions
an accurate physical model for column dynam- ranging from basaltic (e.g., Fontana Lapilli) to
ics that is also computationally fast. rhyolitic (e.g., Taupo) are characterized by larger
maximum thicknesses and more gradual thin-
ning with distance (i.e., larger bt in Eq. 9.2) than
9.6╇ ╇ Case study the 1998 Etna deposit, which instead plots with
two other well-studied subplinian eruptions:
The 22 July 1998 paroxysmal event of Mt. Etna the 17 June 1996 eruption of Ruapehu volcano
provides a useful illustration of the application (New Zealand) and the 22 July 1980 eruption of
of the main empirical, analytical, and numer- Mt. St. Helens (USA). In addition, sensitivity ana-
ical models described in Sections 9.3–9.5 (Coltelli lyses based on time-series Meteosat images and
et€al., 2006; Scollo et al., 2008a). This was one of the theoretical modeling using the PUFF ash tracking
strongest explosive events at Mt. Etna in the last model (Searcy et al., 1998) gave a best-fit value of
century and produced a short-lived strong plume column height of 13 km a.s.l., an eruption dur-
associated with hawaiitic magma (Corsaro and ation between 20 and 40 minutes, and horizontal
Pompilio, 2004) that rose 12 km above sea level and vertical diffusivity values of 5000 and 10 m2
(a.s.l.), ~9 km above the vent in Voragine crater, s−1, respectively (Aloisi et al., 2002; Table 9.1).
Parameter Method
9780521895439c09_p173-202.indd 187
Carey
Power and Wilson and Numerical
Observational Exponential Law Sparks Sparks Walker Inversion: OAT: solution:
Data Method Method (1986) (1986) (1987) TEPHRA2 PUFF FALL3D
Erupted Mass _ 0.9(1seg) 2.0 _ _ _ 1.7 _ 1.7
M (×109 kg) 1.1(2seg)
1.8(3seg)
Column height 12 _ _ 11 _ _ 13 13 12
H (km a.s.l.)
Mass eruption 0.6 (1seg)(8) _ _ _ 0.6(9) 1.8(10) _ _ 2.4
rate Mâ•›̇ (×106 0.7 (2seg)
kg s–1) 1.2 (3seg)
1.3 (PL)
1.1 TEPHRA2
Grainsize Mdϕ _ _ _ _ _ Mdϕ Mdϕ _
distribution (0.8) (-0.6) (0)
STDV STDV STDV
(1.8) (2.2) (1.5)
Duration of 25 (total _ _ _ 24 (1seg) 8 (1seg) _ 20–40 12
sustained duration) 30 (2seg) 10 (2seg)
phase (min) 51 (3seg) 17 (3seg)
56 (PL) 19 (PL)
47 16
TEPHRA2 TEPHRA2
Maximum/ 11 / 6 _ _ 10–30 _ _ 6/6 _ _
average
wind speed
(m s–1)
K H /K V (m2 s–1) _ _ _ _ _ _ 0.5 5000 / 5000 /0.004–
10 600 (mean:
50)
8/2/2012 12:33:41 PM
188 C O S TA N Z A B ON AD ON N A A N D A N TONIO CO STA
Figure 9.5╇╇ Semi-log plot of thickness vs. square root of isopach area for describing the thinning trend of eruptions of different
styles.Vulcanian explosion and dome collapse of Soufrière Hills volcano, Montserrat, West Indies (Bonadonna et al., 2002b).
Subplinian eruptions of Ruapehu volcano, New Zealand (17 June 1996; Bonadonna et al., 2005a); Mount St. Helens, USA (22 July
1980; Sarna-Wojcicki et al., 1981); Kilauea volcano, USA (Keanakakoi tephra, Unit 6; McPhie et al., 1990); Fuego 1974 (Rose et al.,
2008). Plinian eruptions of Askja volcano, Iceland (Unit D; Sparks et al., 1981); Hudson volcano, Chile (12–15 August 1991; Scasso
et al., 1994), Minoan eruption, Greece (Pyle, 1990), Mount St. Helens, USA (18 May 1980; Sarna-Wojcicki et al., 1981); Quizapu,
Chile (1932; Hildreth and Drake, 1992); Tarawera, New Zealand (1886; Walker et al., 1984); Taupo, New Zealand (181AD; Walker,
1980); Masaya volcano, Nicaragua (Fontana Lapilli; Costantini et al., 2008); Masaya volcano, Nicaragua (Triple Layer; Perez et al.,
2009).
sedimentation regimes (break in slope at 15% The goodness of fit is determined as the root
of intermediate-Re particles and 85% of low- mean square error (RMSE):
Re particles). The discrepancy between the
2
observed and computed tephra accumulation N
( Mca − Moa )
is probably due to the under-representation
RMSE = ∑
a =1 Moa
(9.16)
of fine particles in the grainsize distribution
derived from field data, which results from the where N is the number of observations and Moa
dominantly proximal exposure of the deposit. and Mca are, respectively, the observed and com-
As a result, overestimates of terminal veloci- puted deposit (i.e., mass per unit area) at sample
ties will lead to overestimation of tephra accu- location a respectively. In order to illustrate the
mulation (Eq. (9.12); Fig. 9.7). distribution of minimum values of the good-
The TEPHRA2 model can be used in com- ness-of-fit measure, Figure 9.8 shows the RMSE
bination with dedicated inversion techniques corresponding to 0.2-log(mass) increments and
to determine eruption parameters. Connor and 2-km-height increments (Run 2, Table 9.2). The
Connor (2006) applied the downhill simplex wind direction was constrained based on the
method to find the optimized set of eruption dispersal axis indicated by the isomass map
parameters corresponding to a given tephra (Fig. 9.4). Results show that this technique pro-
deposit, based on comparison between observed vides a very good constraint on the erupted mass
and computed mass accumulation per unit area. but not on the column height. The optimized
1000
100
Mass/area (kg/m2)
y = 309.2e–0.71x
10
y = 27.0e–0.25x
1
y = 1.6e–0.07x
0.1
0.01
0 10 20 30 40 50
Area(1/2) (km)
Figure 9.7╇╇ Semi-log plot of mass per unit area (kg m-2) vs.
1000
distance from vent (km) showing the comparison between
Etna 1998 field data (from Figs. 9.4 and 9.5; symbols) and the
100
Mass/area (kg/m2)
Figure 9.6╇╇ Semi-log plots of mass per unit area (kg m−2) fell according to power-law diffusion (Eq. (8) in
vs. square root of area (km) showing (a) the best fit of three Bonadonna et al., 2005b). Figure 9.9 shows the
exponential segments and (b) the power-law fitting. Best-fit forward solution of TEPHRA2 computed using
equations are also shown. these best-fit values and the quantitative com-
parison with field data.
Table 9.2╇ ╇ Parameters used in the simulations. See online supplement 9A for details of the methods used to
derive parameters.
duration and the source terms (i.e., injected of Etna volcano). The turbulent diffusivity ten-
mass per second, column height, mass distribu- sor was described through similarity theory
tion in the plume) were determined through the for the vertical component, and as both con-
model of Bursik (2001), based on BPT. In agree- stant and using a large eddy simulation (LES)
ment with this model, a mass eruption rate Ṁ model for the horizontal component (Folch
of 2.5 × 106 kg s-1 produces a column height of et al., 2009). Note that, when the wind field is
~9 km above the vent (~12 km a.s.l.). Hence, derived from a single sounding, LES models
taking this value for Ṁ and the duration of the can underestimate turbulent diffusion because
climactic phase of 12 minutes (from best-fit ana- there is no horizontal shear. In fact, fits to the
lysis), the total mass erupted is found to be 1.7 × data suggest a constant horizontal diffusion
109 kg. Figure 9.10 shows the deposit obtained of KH ≈ 5000 m2 s−1, compared to the average
by FALL3D using a grid of 51 × 51 × 18 km (1 km value of KH ≈ 2250 m2 s−1 predicted by the LES
horizontal spacing) and the wind field refined model. A comparison of the simulation results
by CALMET. and observed deposit is shown in Fig. 9.10b.
Terminal velocity was calculated using the Finally, FALL3D can also be used to assess ash
model of Ganser (1993) with a sphericity of concentration in the atmosphere and identify
0.93–0.95 (consistent with the data of Coltelli hazardous zones for air traffic (Folch et al.,
et al. (2008) for the 18 December 200s2 eruption 2009).
9.7.2â•…Determination of eruption
parameters
As described above, eruption parameters can
be inferred by applying empirical, analytical,
and numerical models, and through inversion
solutions of analytical models. Our case study
confirms that empirical extrapolation of poor
data sets can be misleading (e.g., volume cal-
culations based on the integration of only one
or two exponential segments), but that integra-
tion of at least three exponential segments and
of the power-law fit give comparable results for
moderate eruptions. However, this result is not
necessarily expected to hold for more exten-
sively dispersed deposits. In addition, sensitiv-
ity tests carried out for the integration of the
power-law fit for the Etna deposit show larger
discrepancies due to the choice of A0 than
of Adist (Table 9.1). This is consistent with the
findings of Bonadonna and Houghton (2005) for
deposits produced by relatively small eruptions
characterized by a power-law exponent m > 2
(e.g., Ruapehu 1996). However, for widely dis-
persed deposits the power-law method is more
sensitive to the choice of Adist than A0 (i.e.,
m < 2). As a result, poorly exposed deposits, in
particular those of large eruptions, would be
better described by applying inversion solutions
of analytical models such as TEPHRA2 (Connor
Figure 9.10╇╇ FALL3D simulation results for the Etna and Connor, 2006), ASHFALL (Hurst and Turner,
22 July 1998 deposit (Run 5, Table 9.2): (a) isomass map 1999), and HAZMAP (Macedonio et al., 2005).
with contours of 0.1, 0.3, 0.4, 0.5, 1, 1.5, 3, 10, 25, and Scollo et al. (2008a) have shown that the model
80 kg m-2, shown for comparison with Figs. 9.4 and 9.9. TEPHRA can be used to determine erupted
Locations of the volcano and the field data are indicated. (b) mass because it is very sensitive to this param-
Comparison between computed (solid line) and observed eter, independent of other inputs, and that the
(symbols) mass/area data. Dashed lines indicate over- or
erupted mass can be well constrained with at
under-estimations of 1/5 and 5 times the observed values
least 10 well-distributed field data points (the
respectively.
data set of our case study has 33 thickness meas-
urements). These conclusions hold for the other
discrepancies in the calculation of mass erup- analytical models. In contrast, determination of
tion rate are due to a combination of factors, column height using TEPHRA is less straight-
including the more complex plume dynamics forward because height is an input, along with
resulting from lava-fountaining fragmentation, the total grainsize distribution, that affects the
and the effects of wind shear on air entrain- model output through interaction with other
ment within the convective region described input parameters (Scollo et al., 2008a). These
in FALL3D, which, for a given plume height, observations are confirmed by application of
requires a larger Ṁ compared to standard ana- inversion techniques using TEPHRA2, which
lytical models. shows that erupted mass is easier to constrain
than column height (e.g., Fig. 9.8). However, probabilistic and/or �real-time forecasting of
Volentik et al. (2010) have shown that a better tephra dispersal. Typically, analytical models
constraint on column height is obtained when are computationally fast and therefore can be
tephra accumulation is inverted for individual used to compile fully probabilistic assessments,
grainsizes. Further studies of the application whereas numerical models are more suited
of inversion techniques are needed to assess for real-time forecasting to provide accur-
the minimum amount of field data required to ate estimates of ground sedimentation and of
provide a reliable characterization. In addition, the position of the volcanic cloud with time.
inversion techniques are not always straightfor- In particular, our case study confirms that the
ward to apply; the choice of parameter ranges more sophisticated numerical models do not
can significantly affect the final result and there- necessarily provide better accuracy in terms
fore requires critical analysis. We also stress that of ground sedimentation, especially given the
studies of poorly exposed deposits have shown uncertainties associated with the input param-
that the method of Carey and Sparks (1986) gives eters (e.g., Scollo et al., 2008b; Figs. 9.9 and
good results even when the position of the vent is 9.10). However, numerical models can provide
not well known (e.g., Wehrmann et al., 2006). As crucial information not possible with analytical
a result, inversion techniques should always be models (e.g., a description of cloud movement
used in combination with other models for con- with time). In general, verified and validated
straining both erupted mass and plume height. numerical models are also more appropriate
In particular, granulometry data in proximal, for real-time forecasting because they require
medial, and distal areas should always be col- fewer empirical parameters (e.g., diffusion
lected to better constrain plume height. Finally, coefficient, mass distribution within the erup-
models based on empirical observations should tive column) and, as a result, are simpler to
always be applied in their range of validation. apply. Models of all levels of sophistication
would benefit significantly from better param-
eterization of critical sedimentation processes
9.8╇ ╇ Summary and outlook such as particle aggregation and settling, and
from quantification of uncertainties associated
Comprehensive characterization of tephra with input parameters such as erupted mass,
deposits and reliable hazard assessment can plume height, mass discharge rate, and total
only result from critical and synergistic appli- grainsize distribution.
cation of models with levels of sophistication
ranging from purely empirical to fully numer-
ical. First, tephra deposits need to be sampled 9.9╇ ╇ Notation
accurately (for both mass/area and grain-size
data) over an area proportional to the associated a sample location in application of
particle dispersal. Second, dedicated empirical inversion techniques
and analytical models can be used for deter- A area enclosed within isomass/isopach
mination of plume height, erupted mass, ini- contours (m2)
tial grainsize distribution, mass eruption rate Ax area enclosed within isopach line of
and duration. Inversion solutions of analytical thickness Tx (m2)
models can also be used to obtain independ- A0 integration limit of power-law
ent results for the same eruption parameters. function, typically taken as distance
Eruption parameters can be used to classify of maximum deposit thickness (m)
volcanic eruptions and build potential eruptive Adist integration limit of power-law
scenarios. Finally, following thorough model function, typically taken as the
validation and calibration, analytical and downwind extent of deposit (m)
numerical models can be applied to compile bt thickness half distance (m)
cj(x,y,z,t) particle concentration (kg m−3) S(x,y,z,t) source term (kg m−3 s−1)
c′j turbulent fluctuation of concentration t time coordinate (s)
of particle class j (kg m-3) T thickness of tephra deposit (m)
c̅j ensemble average of concentration Tn0 intercept of exponential segment n
of particle class j (kg m-3) (m)
C empirical factor (m kg-1/4 s1/4) T0 maximum thickness of tephra
CD drag coefficient deposit (m)
Cpl coefficient of power-law best fit Tx thickness of given isopach x (m)
(m(1+m)) ux,uy,uz components of wind velocity vector
dj diameter of particle of class j (m) (m s−1)
F particle shape factor u̅x,u̅y,u̅z components of average wind velocity
H maximum height of volcanic plume (m s−1)
(m) u′x,u′y,u′z components of turbulent fluctuation
Hb neutral buoyancy level of volcanic of wind velocity (m s−1)
plume (m) U current velocity in s-direction of
j index of particle size plume or umbrella cloud (m s−1)
k slope of deposit exponential best-fit vHb particle terminal velocity at neutral
curve (m-1) buoyancy level Hb (m s−1)
kn slope of exponential segment n (m−1) vj particle terminal velocity of particle
KH horizontal atmospheric diffusion class j; vj = vj (x,y,z) (m s−1)
coefficient (KH = Kx = Ky) (m2 s−1) V erupted volume (m3)
Kx x-component (horizontal) of w maximum crosswind width at source
horizontal diffusion coefficient of spreading current (m)
(m2 s−1) x o plume corner position (m)
K V vertical atmospheric diffusion x, y, z spatial coordinates (m)
coefficient (m2 s−1) ηa dynamic viscosity of air (Pa s)
Ky y-component (horizontal) of ρa density of air (kg m−3)
horizontal diffusion coefficient (m2 s−1) ψ sphericity
m exponent of power-law best fit curve ψwork sphericity
L, I, S dimensions of three perpendicular Re Reynolds number
axes of particle (m)
M total mass of particles of given
size fraction carried by current at Acknowledgments
distance x1 (kg)
Ṁ mass eruption rate (kg s−1) A. Costa was supported by the MIUR-FIRB
M0 initial mass for given grain size Italian project “Sviluppo Nuove Tecnologie
injected into current (kg) per la Protezione e Difesa del Territorio dai
Moa observed mass per unit area at Rischi Naturali.” The authors are grateful to
sample location a (kg m−2) D. Andronico, M. Coltelli, and P. Del Carlo for
Mca computed mass per unit area at providing the field data of the 22 July 1998 erup-
sample location a (kg m−2) tion of Mt. Etna and to A. Folch for providing the
n number of exponential segments pseudo-sounding used to run CALMET. S. Scollo,
N number of field observations in and L. Connor are thanked for useful discussion.
application of inversion techniques
Q volumetric flow rate into current at
neutral buoyancy level (m3 s−1) References
s spreading direction of current (m)
Sn position of break in slope of Aloisi, M., D’Agostino, M., Dean, K. G., Mostaccio,
exponential segment n (m) A. and Neri, G. (2002). Satellite analysis and
PUFF simulation of the eruptive cloud generated Bonadonna, C., Macedonio, G. and Sparks, R. S. J.
by the Mount Etna paroxysm of 22 July 1998. (2002a). Numerical modelling of tephra fallout
Journal of Geophysical Research, 107(B12), 2373, associated with dome collapses and Vulcanian
doi:10.1029/2001JB000630. explosions: application to hazard assessment on
Andronico, D, Del Carlo, P. and Coltelli, M. (1999). Montserrat. In The Eruption of Soufrière Hills Volcano,
The 22 July 1998 fire fountain episode at Voragine Montserrat, from 1995 to 1999, ed. T. H. Druitt and B.
Crater (Mt. Etna, Italy). Volcanic and Magmatic Studies P. Kokelaar : Geological Society London Memoir,
Group Annual Meeting, UK, 5–6 January. 517–537.
Andronico, D., Scollo, S., Caruso, S. and Cristaldi, Bonadonna, C., Mayberry, G. C., Calder, E. et al.
A. (2008). The 2002–03 Etna explosive activity: (2002b). Tephra fallout in the eruption of Soufrière
Tephra dispersal and features of the deposits. Hills Volcano, Montserrat. In The Eruption of
Journal of Geophysical Research, 113, B04209, Soufrière Hills Volcano, Montserrat, from 1995 to 1999,
doi:10.1029/2007JB005126. ed. T. H. Druitt and B. P. Kokelaar : Geological
Arastoopour, H., Wang, C. H. and Weil, S. A. (1982). Society London Memoir, 483–516.
Particle-particle interaction force in a diluite Bonadonna, C., Connor, C. B., Houghton, B. F.
gas-solid system. Chemical Engineering Sciences, 37, et al. (2005b). Probabilistic modeling of
1379–1386. tephra dispersion: hazard assessment of a
Armienti, P., Macedonio, G. and Pareschi, M. T. multi-phase eruption at Tarawera, New Zealand.
(1988). A numerical model for simulation of tephra Journal of Geophysical Research, 110, B03203,
transport and deposition€– applications to May doi:10.1029/2003JB002896.
18, 1980, Mount St. Helens eruption. Journal of Bonadonna, C., Phillips, J. C. and Houghton, B. F.
Geophysical Research, 93(B6), 6463–6476. (2005a). Modeling tephra sedimentation
Aschenbrenner, B. C. (1956). A new method of from a Ruapehu weak plume eruption.
expressing particle sphericity. Journal of Sedimentary Journal of Geophysical Research, 110, B08209,
Petrology, 26, 15–31. doi:10.1029/2004JB003515.
Barberi, F., Coltelli, M., Frullani, A., Rosi, M. and Brazier, S., Davis, A. N., Sigurdsson, H. and Sparks,
Almeida, E. (1995). Chronology and dispersal R. S. J. (1982). Fallout and deposition of volcanic
characteristics of recently (last 5000 years) erupted ash during the 1979 explosive eruption of the
tephra of Cotopaxi (Ecuador): implications for Soufrière of St. Vincent. Journal of Volcanology and
long-term eruptive forecasting. Journal of Volcanology Geothermal Research, 14, 335–359.
and Geothermal Research, 69, 217–239. Bursik, M. (2001). Effect of wind on the rise height of
Barsotti, S. and Neri, A. (2008). The VOL-CALPUFF volcanic plumes. Geophysical Research Letter, 28(18),
model for atmospheric ash dispersal: 2. 3621–3624.
Application to the weak Mount Etna plume of July Bursik, M. I., Carey, S. N. and Sparks, R. S. J. (1992b).
2001. Journal of Geophysical Research, 113, B03209, A gravity current model for the May 18, 1980
doi:10.1029/2006JB004624. Mount-St-Helens plume. Geophysical Research Letters,
Barsotti, S., Neri, A. and Scire, J. S. (2008). The 19, 1663–1666.
VOL-CALPUFF model for atmospheric ash Bursik, M. I., Sparks, R. S. J., Gilbert, J. S. and
dispersal: 1. Approach and physical formulation. Carey, S. N. (1992a). Sedimentation of tephra by
Journal of Geophysical Research, 113, B03208, volcanic plumes: I. Theory and its comparison
doi:10.1029/2006JB004623. with a study of the Fogo A plinian deposit, Sao
Bonadonna, C. and Phillips, J. C. (2003). Miguel (Azores). Bulletin of Volcanology, 54, 329–344.
Sedimentation from strong volcanic plumes. Byrne, M. A., Laing, A. G. and Connor, C. (2007).
Journal of Geophysical Research, 108(B7), 2340–2368. Predicting tephra dispersion with a mesoscale
Bonadonna, C. and Houghton, B. F. (2005). Total atmospheric model and a particle fall model:
grainsize distribution and volume of tephra-fall Application to Cerro Negro volcano. Journal of
deposits. Bulletin of Volcanology, 67, 441–456. Applied Meteorology and Climatology, 46, 121–135.
Bonadonna, C., Ernst, G. G. J. and Sparks, R. S. J. Carazzo, G., Kaminski, E. and Tait, S. (2008). On the
(1998). Thickness variations and volume estimates dynamics of volcanic columns: A comparison of
of tephra fall deposits: the importance of particle field data with a new model of negatively buoyant
Reynolds number. Journal of Volcanology and jets. Journal of Volcanology and Geothermal Research,
Geothermal Research, 81, 173–187. 178, 94–103.
Carey, S. N. and Sigurdsson, H. (1982). Influence Journal of Volcanology and Geothermal Research, 137,
of particle aggregation on deposition of distal 55–71.
tephra from the May 18, 1980, eruption of Mount Costa, A., Macedonio, G. and Folch, A. (2006). A
St-Helens volcano. Journal of Geophysical Research, three-dimensional Eulerian model for transport
87(B8), 7061–7072. and deposition of volcanic ashes. Earth and
Carey, S. N. and Sigurdsson, H. (1986). The 1982 Planetary Science Letters, 241, 634–647.
eruptions of El Chichon volcano, Mexico (2): Costantini, L., Bonadonna, C., Houghton, B. F.
observations and numerical modelling of and Wehrmann, H. (2008). New physical
tephra-fall distribution. Bulletin of Volcanology, 48, characterization of the Fontana Lapilli basaltic
127–141. Plinian eruption, Nicaragua. Bulletin of Volcanology,
Carey, S. and Sigurdsson, H. (1989). The intensity of 71, 337–355.
Plinian eruptions. Bulletin of Volcanology, 51, 28–40. D’amours, R. (1998). Modeling the ETEX plume
Carey, S. N. and Sparks, R. S. J. (1986). Quantitative dispersion with the Canadian emergency response
models of the fallout and dispersal of tephra from model. Atmospheric Environment, 32, 4335–4341.
volcanic eruption columns. Bulletin of Volcanology, Dobran, F., Neri, A. and Macedonio, G. (1993).
48, 109–125. Numerical simulation of collapsing volcanic
Chhabra, R. P., Agarwal, L. and Sinha, N. K. (1999). columns. Journal of Geophysical Research, 98,
Drag on non-spherical particles: an evaluation of 4231–4259.
available methods. Powder Technology, 101, 288–295. Draxier, R. R. and Hess, G. D. (1998). An overview of
Cole, P. D., Calder, E. S., Druitt, T. H. et al. (1998). the HYSPLIT_4 modelling system for trajectories,
Pyroclastic flows generated by gravitational dispersion and deposition. Australian Meteorological
instability of the 1996–97 lava dome of Soufrière Magazine, 47, 295–308.
Hills Volcano, Montserrat. Geophysical Research Durant, A. J., Rose, W. I., Sarna-Wojcicki, A. M.,
Letters, 25, 3425–3428. Carey, S. and Volentik, A. C. M. (2009).
Coltelli, M., Puglisi, G., Guglielmino, F. and Hydrometeor-enhanced tephra sedimentation:
Palano, M. (2006). Application of differential SAR Constraints from the 18 May 1980 eruption of
interferometry for studying eruptive event of 22 Mount St. Helens. Journal of Geophysical Research,
July 1998 at Mt. Etna. Quaderni di Geofisica, 43, 114, B03204, doi:10.1029/2008JB005756.
15–20. Esposti Ongaro, T., Cavazzoni, C., Erbacci, G., Neri, A.
Coltelli, M., Miraglia, L. and Scollo, S. (2008). and Salvetti, M. V. (2007). A parallel multiphase
Characterization of shape and terminal velocity of flow code for the 3D simulation of explosive
tephra particles eurpted during the 2002 eruption volcanic eruptions. Parallel Computing, 33, 541–560.
of Etna volcano, Italy. Bulletin of Volcanology, 70, Fierstein, J. and Nathenson, M. (1992). Another
1103–1112. look at the calculation of fallout tephra volumes.
Connor, L. G. and Connor, C. B. (2006). Inversion is Bulletin of Volcanology, 54, 156–167.
the key to dispersion: understanding eruption Folch, A. and Felpeto, A. (2005). A coupled model
dynamics by inverting tephra fallout. In Statistics in for dispersal of tephra during sustained explosive
Volcanology, ed. H. Mader, S. Cole, C. B. Connor and eruptions. Journal of Volcanology and Geothermal
L. G. Connor. London: Geological Society 231–242. Research, 145, 337–349.
Connor, C. B., Hill, B. E., Winfrey, B., Franklin, N. M. Folch, A., Jorba, O. and Viramonte, J. (2008). Volcanic
and La Femina, P. C. (2001). Estimation of volcanic ash forecast€– application to the May 2008 Chaiten
hazards from tephra fallout. Natural Hazards Review, eruption. Natural Hazards and Earth System Sciences,
2, 33–42. 8, 927–940.
Cornell, W., Carey, S. and Sigurdsson, H. (1983). Folch, A., Costa, A. and Macedonio, G. (2009).
Computer-simulation of transport and deposition FALL3D: A computational model for transport
of the Campanian Y-5 Ash. Journal of Volcanology and and deposition of volcanic ash. Computers and
Geothermal Research, 17, 89–109. Geosciences, 35, 1334–1342
Corsaro, R. A. and Pompilio, M. (2004). Magma Froggatt, P. C. (1982). Review of methods estimating
dynamics in the shallow plumbing system of rhyolitic tephra volumes; applications to the Taupo
Mt. Etna as recorded by compositional variations Volcanic Zone, New Zealand. Journal of Volcanology
in volcanics of recent summit activity (1995–1999). and Geothermal Research, 14, 1–56.
Ganser, G. H. (1993). A rational approach to drag sequence in explosive volcanic eruptions. Journal of
prediction of spherical and nonspherical particles. Geophysical Research, 103(B12), 29,759–29,779.
Powder Technology, 77, 143–152. Koyaguchi, T. (1994). Grain-size variation of tephra
Glaze, L. S. and Self, S. (1991). Ashfall dispersal derived from volcanic umbrella clouds. Bulletin of
for the 16 September 1986, eruption of Lascar, Volcanology, 56, 1–9.
Chile, calculated by a turbulent-diffusion model. Koyaguchi, T. and Ohno, M. (2001a). Reconstruction
Geophysical Research Letters, 18, 1237–1240. of eruption column dynamics on the basis of grain
Glaze, L. S., Wilson, L. and Mouginis-Mark, P. J. size of tephra fall deposits. 1. Methods. Journal of
(1999). Volcanic eruption plume top topography Geophysical Research, 106(B4), 6499–6512.
and heights as determined from photoclinometric Koyaguchi, T. and Ohno, M. (2001b). Reconstruction
analysis of satellite data. Journal of Geophysical of eruption column dynamics on the basis of
Research, 104(B2), 2989–3001. grain size of tephra fall deposits. 2. Application to
Heffter, J. L. and Stunder, B. J. B. (1993). Volcanic the Pinatubo 1991 euption. Journal of Geophysical
Ash Forecast Transport and Dispersion (VAFTAD) Research, 106(B4), 6513–6533.
Model. Weather and Forecasting, 8, 533–541. Kunii, D. and Levenspiel, O. (1969). Fluidisation
Herzog, M., Graf, H. F., Textor, C. and Oberhuber, Engineering. New York: Wiley and Sons.
J. M. (1998). The effect of phase changes of water Legros, F. (2000). Minimum volume of a tephra
on the development of volcanic plumes. Journal of fallout deposit estimated from a single isopach.
Volcanology and Geothermal Research, 87, 55–74. Journal of Volcanology and Geothermal Resarch, 96,
Hildreth, W. and Drake, R. E. (1992). Volcano Quizapu, 25–32.
Chilean Andes. Bulletin of Volcanology, 54, 93–125. Macedonio, G., Costa, A. and Longo, A. (2005). A
Hobbs, P. V., Lyons, J. H., Locatelli, J. D. et al. (1981). computer model for volcanic ash fallout and
Radar detection of cloud-seeding effects. Science, assessment of subsequent hazard. Computers and
213, 1250–1252. Geosciences, 31, 837–845.
Holasek, R. E. and Self, S. (1995). GOES Mannen, K. (2006). Total grain size distribution of
weather-satellite observations and measurements a mafic subplinian tephra, TB-2, from the 1986
of the May 18, 1980, Mount St. Helens Eruption. Izu-Oshima eruption, Japan: An estimation based
Journal of Geophysical Research, 100(B5), 8469–8487. on a theoretical model of tephra dispersal. Journal
Hurst, A. W. and Turner, R. (1999). Performance of Volcanology and Geothermal Research, 155, 1–17.
of the program ASHFALL for forecasting ashfall Martin, D. and Nokes, R. (1988). Crystal settling in
during the 1995 and 1996 eruptions of Ruapehu a vigorously convecting magma chamber. Nature,
volcano. New Zealand Journal of Geology and 332, 534–536.
Geophysics, 42, 615–622. McPhie, J., Walker, G. P. L. and Christiansen, R. L.
Ishimine, Y. (2006). Sensitivity of the dynamics of (1990). Phreatomagmatic and phreatic fall and
volcanic eruption columns to their shape. Bulletin surge deposits from explosions at Kilauea volcano,
of Volcanology, 68(6), 516–537. Hawaii, 1790 A.D.: Keanakakoi Ash Member.
James, M. R., Gilbert, J. S. and Lane, S. J. (2002). Bulletin of Volcanology, 52, 334–354.
Experimental investigation of volcanic particle Michalakes, J., Dudhia, J., Gill, D. et al. (2005). The
aggregation in the absence of a liquid phase. weather research and forecast model: Software
Journal of Geophysical Research, 107(B9), 2191, architecture and performance. Use of High
doi:10.1029/2001JB000950. Performance Computing in Meteorology, Proceedings of
James, M. R., Lane, S. J. and Gilbert, J. S. (2003). the Eleventh ECMWF Workshop, 156–168.
Density, construction, and drag coefficient Morton, B., Taylor, G. L. and Turner, J. S. (1956).
of electrostatic volcanic ash aggregates. Turbulent gravitational convection from
Journal of Geophysical Research, 108(B9), 2435, maintained and instantaneous source. Proceedings
doi:10.1029/2002JB002011. of the Royal Society of London, A234, 1–23.
Kalnay, E., Kanamitsu, M., Kistler, R. et al. (1996). The Neri, A., Papale, P. and Macedonio, G. (1998). The
NCEP/NCAR 40-year reanalysis project. Bulletin of role of magma composition and water content
the American Meteorological Society, 77, 437–471. in explosive eruptions: 2. Pyroclastic dispersion
Kaminski, E. and Jaupart, C. (1998). The size dynamics. Journal of Volcanology and Geothermal
distribution of pyroclasts and the fragmentation Research, 87, 95–115.
Oberhuber, J. M., Herzog, M., Graf, H. F. and Ryall, D. B. and Maryon, R. H. (1998). Validation of
Schwanke, K. (1998). Volcanic plume simulation the UK Met. Office’s name model against the ETEX
on large scales. Journal of Volcanology and Geothermal dataset. Atmospheric Environment, 32, 4265–4276.
Research, 87, 29–53. Sandu, I., Bompay, F. and Stefan, S. (2003). Validation
Papale, P. and Rosi, M. (1993). A case of no-wind of atmospheric dispersion models using ETEX data.
plinian fallout at Pululagua caldera (Ecuador): International Journal of Environment and Pollution, 19,
implications for model of clast dispersal. Bulletin of 367–389.
Volcanology, 55, 523–535. Sarna-Wojcicki, A. M., Shipley, S., Waitt, J. R.,
Perez, W., Freundt, A., Kutterolf, S. and Schmincke, Dzurisin, D. and Wood, S. H. (1981). Areal
H. U. (2009). The Masaya Triple Layer: A 2100 year distribution thickness, mass, volume, and
old basaltic multi-episodic Plinian eruption from grain-size of airfall ash from the six major
the Masaya Caldera Complex (Nicaragua). Journal of eruptions of 1980. In The 1980 Eruption of Mount
Volcanology and Geothermal Research, 179, 191–205. St. Helens, ed. W. P. Lipman and D. R. Mullineaux.
Pfeiffer, T., Costa, A. and Macedonio, G. (2005). A Washington, D.C.: U.S. Geological Survey
model for the numerical simulation of tephra Professional Paper, 1250, 577–600.
fall deposits. Journal of Volcanology and Geothermal Scasso, R., Corbella, H. and Tiberi, P. (1994).
Resarch, 140, 273–294. Sedimentological analysis of the tephra from
Prata, A. J. and Grant, I. F. (2001). Retrieval of 12–15 August 1991 eruption of Hudson Volcano.
microphysical and morphological properties Bulletin of Volcanology, 56, 121–132.
of volcanic ash plumes from satellite data: Scire, J. S., Robe, F. and Yamartino, R. (2000). A User’s
Application to Mt Ruapehu, New Zealand. Quarterly Guide for the CALMET Meteorological Model. Concord,
Journal of the Royal Meteorological Society, 127, MA: Earth Tech, Inc.
2153–2179. Scollo, S., Del Carlo, P. and Coltelli, M. (2007). Tephra
Prata, A. J. and Turner, P. J. (1997). Cloud-top height fallout of 2001 Etna flank eruption: Analysis of the
determination using ATSR data. Remote Sensing of deposit and plume dispersion. Journal of Volcanology
Environment, 59, 1–13. and Geothermal Research, 160, 147–164.
Pyle, D. M. (1989). The thickness, volume and Scollo, S., Folch, A. and Costa, A. (2008b). A
grainsize of tephra fall deposits. Bulletin of parametric and comparative study of different
Volcanology, 51, 1–15. tephra fallout models. Journal of Volcanology and
Pyle, D. M. (1990). New estimates for the volume of Geothermal Research, 176, 199–211.
the Minoan eruption. In Thera and the Aegean World, Scollo, S., Tarantola, S., Bonadonna, C., Coltelli, M.
ed. D. A. Hardy. London: The Thera Foundation, and Saltelli, A. (2008a). Sensitivity analysis and
113–121. uncertanity estimation for tephra dispersal
Pyle, D. M. (1995). Assessment of the minimum models. Journal of Geophysical Research, 113, B06202,
volume of tephra fall deposits. Journal of Volcanology doi:10.1029/2006JB004864.
and Geothermal Research, 69, 379–382. Searcy, C., Dean, K. and Stringer, W. (1998). PUFF: A
Riley, C. M., Rose, W. I. and Bluth, G. J. S. (2003). high-resolution volcanic ash tracking model. Journal
Quantitative shape measurements of distal of Volcanology and Geothermal Research, 80, 1–16.
volcanic ash. Journal of Geophysical Research, 108, Settle, M. (1978). Volcanic eruption clouds and
2504, doi:10.1029/2001JB000818. thermal power output of explosive eruptions.
Rose, W. I. (1993). Comment on “Another look at the Journal of Volcanological and Geothermal Research, 3,
calculation of fallout tephra volumes”. Bulletin of 309–324.
Volcanology, 55, 372–374. Shaw, D. M., Watkins, N. D. and Huang, T. C. (1974).
Rose, W. I., Self, S., Murrow, P. J. et al. (2008). Nature Atmospherically transported volcanic glass in
and significance of small volume fall deposits at deep-sea sediments: Theoretical considerations.
composite volcanoes: Insights from the October Journal of Geophysical Research, 79, 3087–3094.
14, 1974 Fuego eruption, Guatemala. Bulletin of Skamarock, W., Klemp, J., Dudhia, J. et al. (2005) A
Volcanology, 70, 1043–1067. description of the Advanced Research WRF Version 2.
Rosi, M. (1998). Plinian eruption columns: particle Available online at: www.wrf- model.org).
transport and fallout. In From Magma to Tephra: Sorem, R. K. (1982). Volcanic ash clusters: tephra
Modelling Physical Processes of Explosive Volcanic Eruptions, rafts and scavengers. Journal of Volcanology and
ed. A. Freundt and M. Rosi. Elsevier, pp. 139–172. Geothermal Research, 13, 63–71.
Sparks, R. S. J. (1986). The dimensions and dynamics Wadell, H. (1933). Sphericity and roundness of rock
of volcanic eruption columns. Bulletin of Volcanology, particles. Journal of Geology, 41, 310–331.
48, 3–15. Walker, G. P. L. (1973). Explosive volcanic eruptions€–
Sparks, R. S. J., Wilson, L. and Sigurdsson, H. (1981). a new classification scheme. Geologische Rundschau,
The pyroclastic deposits of the 1875 eruption of 62, 431–446.
Askja, Iceland. Philosophical Transactions of the Royal Walker, G. P. L. (1980). The Taupo Pumice: product of
Society of London, 229, 241–273. the most powerful known (Ultraplinian) eruption?
Sparks, R. S. J., Carey, S. N. and Sigurdsson, H. (1991). Journal of Volcanology and Geothermal Research, 8,
Sedimentation from gravity currents generated by 69–94.
turbulent plumes. Sedimentology, 38, 839–856. Walker, G. P. L., Wilson, L. and Bowell, E. L. G.
Sparks, R. S. J., Bursik, M. I., Ablay, G. J., Thomas, (1971). Explosive volcanic eruptions€– I. The rate
R. M. E. and Carey, S. N. (1992). Sedimentation of fall of pyroclasts. Geophysical Journal of the Royal
of tephra by volcanic plumes. 2. Controls on Astronomical Society, 22, 377–383.
thickness and grain-size variations of tephra fall Walker, G. P. L., Self, S. and Wilson, L. (1984).
deposits. Bulletin of Volcanology, 54, 685–695. Tarawera, 1886, New Zealand€– A basaltic
Sparks, R. S. J., Bursik, M. I., Carey, S. N. et al. (1997). Plinian fissure eruption. Journal of Volcanology and
Volcanic Plumes. Chichester, UK: Wiley. Geothermal Research, 21, 61–78.
Stull, R. (1988). An Introduction to Boundary Layer Wehrmann, H., Bonadonna, C., Freundt, A.,
Meteorology. Dordrecht: Kluwer Academic. Houghton, B. F. and Kutterolf, S. (2006). Fontana
Sulpizio, R. (2005). Three empirical methods for the Tephra: A basaltic Plinian eruption in Nicaragua.
calculation of distal volume of tephra-fall deposits. In Volcanic Hazards in Central America, Geological
Journal of Volcanology and Geothermal Research, 145, Society of America Special Paper, 412, pp.
315–336. 209–223.
Suzuki, T. (1983). A theoretical model for dispersion Wiesner, M. G., Wang, Y. W. and Zheng, L. (1995).
of tephra. In Arc Volcanism, Physics and Tectonics, Fallout of volcanic ash to the deep South China Sea
ed. D. Shimozuru and I. Yokoyama. Tokyo: Terra induced by the 1991 eruption of Mount Pinatubo
Scientific, pp. 95–113. (Philippines). Geology, 23, 885–888.
Tanaka, H. L. and Yamamoto, K. (2002). Numerical Wilson, L. and Huang, T. C. (1979). The influence
simulation of volcanic plume dispersal from Usu of shape on the atmospheric settling velocity of
volcano in Japan on 31 March 2000 using PUFF volcanic ash particles. Earth and Planetary Sciences
model. Earth Planets Space, 54, 743–752. Letters, 44, 311–324.
Textor, C., Graf, H. F., Herzog, M. et al. (2006). Wilson, L. and Walker, G. P. L. (1987). Explosive
Volcanic particle aggregation in explosive volcanic eruptions€– VI. Ejecta dispersal in
eruption columns. Part II: Numerical experiments. plinian eruptions€– the control of eruption
Journal of Volcanology and Geothermal Research, 150, conditions and atmospheric properties.
378–394. Geophysical Journal of the Royal Astronomical Society,
Thorarinsson, S. (1944). Petrokronologista Studier pa 89, 657–679.
Island. Geographes Annuales Stockholm, 26, 1–217. Wilson, L., Sparks, R. S. J., Huang, T. C. and Watkins,
Thorarinsson, S. (1954). The eruption of Hekla N. D. (1978). The control of volcanic column height
1947–1948. In The Tephra Fall from Hekla. Reykjavik: by eruption energetics and dynamics. Journal of
Vis Islendinga. Geophysical Research, 83, 1829–1836.
Turner, J. S. (1979). Buoyancy Effects in Fluids. Witham, C. S., Hort, M. C., Potts, R. et al. (2007).
Cambridge: Cambridge University Press. Comparison of VAAC atmospheric dispersion
Veitch, G. and Woods, A. W. (2001). Particle models using the 1 November 2004 Grimsvotn
aggregation in volcanic eruption columns. Journal eruption. Meteorological Applications, 14, 27–38.
of Geophysical Research, 106(B11), 26,425–26,441. Woods, A. W. (1988). The fluid dynamics and
Volentik, A., Bonadonna, C., Connor, C. B., Connor, thermodynamics of eruption columns. Bulletin of
L. J. and Rosi, M. (2010). Modeling tephra dispersal Volcanology, 50, 169–193.
in absence of wind: insights from the climactic Zimanowski, B., Wohletz, K., Dellino, P. and
phase of the 2450BP Plinian eruption of Pululagua Buttner, R. (2003). The volcanic ash problem.
volcano (Ecuador). Journal of Volcanology and Journal of Volcanology and Geothermal Research,
Geothermal Research, 193, 117–136. 122, 1–5.
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
Table 10.1╇ ╇ Selected microscopic and macroscopic parameters, and dimensionless numbers relevant to pyroclastic density
currents.
solid volume fraction can be high and from deposits that are commonly poorly sorted and
which deposition occurs. Surges are weakly massive. The nature of their internal velocity
influenced by topography as they are tens to profile is poorly constrained. Two conceptual
hundreds of meters thick, and their deposits are models for the deposition mechanisms of PDCs
commonly well sorted and laminated. On the close to the dense end member have been pro-
other hand, dense flows (e.g., pumice-and-ash, posed: en masse deposition, in which the current
block-and-ash, and scoria flows) have particle freezes through its entire height (Sparks, 1976),
concentrations of the same order of magni- and progressive aggradation in which the deposit
tude as those of their deposits. They are a few builds up by accumulation of material (Fisher,
meters to a few tens of meters thick, topograph- 1966; Branney and Kokelaar, 1992). The dense
ically controlled, pond in depressions, and form and dilute end members can coexist in most
PDCs, consisting of a coupled concentrated basal result whenever fluid of one density flows hori-
avalanche (the flow) and an overriding ash cloud zontally into a fluid of a different density, and
(the surge) (Fig. 10.1), which results in complex are frequent occurrences in the natural world.
internal particle concentration and velocity pro- Dust storms and sea breeze fronts, thunder-
files through the depth of the current. Note also storms and estuarine outflows, deep ocean tur-
that a single eruption may produce PDCs that bidity flows and PDCs are just a few examples
vary in a continuum between the end members. (Simpson, 1997). In these cases, the contrast in
When PDCs interact with topography, they can density between the two fluids can arise from
incorporate ambient air whereas the overriding compositional or thermal differences between
ash-cloud surge may generate secondary dense the fluids, or the presence of suspended parti-
flows as it sediments particles out of suspension cles of a different density. The dynamics of grav-
(Druitt et€al., 2002). ity currents also depend strongly on whether
In the context of volcanic hazard assess- the fluid that forms the current is released as a
ment and mitigation, predictive theoretical finite volume or continuously from a sustained
models require improved understanding of the source. An important paradigm for gravity cur-
complex physics of PDCs, and experimental rent initiation is the gravitational collapse of a
laboratory investigations can help provide fun- column of dense fluid or particles (Figs. 10.1(c),
damental insights. Dynamic similarity between 10.1(d), and 10.2); recent experimental studies
the large-scale PDCs and their small-scale ana- have led to the recognition that the dynamics
logs is ensured when the dynamically relevant of the collapse of a column of particles in air
dimensional groupings are equivalent (see shares many common features with the fluid
Table 10.1). In this chapter, we first describe case (Lube et€ al. 2004, Balmforth and Kerswell,
the basic physics of gravity-driven motion of 2005; Lajeunesse et€al., 2004; Roche et€al., 2008).
particle suspensions, and introduce simplified The dynamics of gravity currents are now
equations of motion for these currents. We generally well understood for a wide range of
then highlight important insights that have conditions (Simpson, 1997; Huppert, 2006).
been obtained from laboratory experiments In this section, we introduce the fundamental
on the gravitational collapse of dense fluid sus- dynamics of gravity currents, focusing in par-
pensions, initially gas-fluidized particles, and ticular on the underlying physics and dynamical
granular columns, emphasizing the distinct aspects that are relevant to PDCs. We start by
effect on the frictional resistance to motion of considering the motion that develops between
these different initial states. We then describe two fluids that have a small and constant density
numerical methods for solving the equations contrast, using energy balances to develop scal-
of motion relevant to emplacement following ing relationships and simple theoretical models
gravitational collapse, with particular focus on for gravity current motion. These approaches
how to combine these dynamical models with are then extended to apply to dilute turbulent
complex natural topography for volcanic haz- PDCs (i.e., high Reynolds number, see Table
ard prediction. Finally, we identify current and 10.1), where the density contrast varies during
future research questions. the motion of the current, as a consequence
of the sedimentation of particles initially held
in suspension by turbulence (Valentine, 1987;
Dellino et€al., 2008). We also consider situations
10.2╇ ╇Fundamental physics in which the density contrast between the two
of gravity currents fluids is large (i.e., dense PDCs), including granu-
lar flows, where the interstitial fluid phase can
10.2.1â•… Generalities influence the flow dynamics to varying extents.
PDCs are an example of a predominantly hori- Figure 10.2 shows a series of images of the
zontal, gravity-driven flow, more generally collapse of a column of a dense fluid and of ini-
known as a gravity current. Gravity currents tially gas-fluidized particles in a channel. The
in the region of initial collapse, and The sedimentation from a flow consisting of
2/3 2/3 a deep turbulent layer above a narrow basal
3 2
Fr 2 / 3 ( g ′ Q 0 )
1/ 3
L= t ( λ + 2)/ 3 (10.5) layer (where deposition can take place) can be
2 λ + 2
described by an approximate particle settling
for subsequent motion where inertia and buoy- law, sometimes known as Hazen’s Law, and
ancy are in balance. For the special case of experimentally verified by Martin and Nokes
motion in deep surrounding fluid following (1988), of the form
a finite volume release, the propagation of a
u t
gravity current with constant density can be c = c0 exp − s , (10.8)
hc
written,
where c is the volume concentration of particles
L ≈ 1.47( g ′ Q 0 )1/ 3 t 2 / 3 . (10.6) suspended in the current with initial value c0. To
complete the model, one further relationship
is required to describe the change in current
10.2.4â•…Fluid gravity currents with density with particle concentration. In the case
particles where the excess density in the current arises
In many natural situations, including PDCs, solely from the presence of the particles (i.e.,
the density difference between gravity currents the density of the current fluid is equal to the
and the surrounding fluid depends on the con- density of the surrounding fluid), we can write
centration of suspended dense particles within
g (ρ p − ρ0 )c
the current. In this case, the bulk density of the g′ = = g ′ p c, (10.9)
current will vary in time if particles are sedi- ρ0
mented at the current base. The framework of where ρp is the particle density. The solution
box models (or more sophisticated formulations for the system of equations (10.2), (10.3), (10.8),
of the conservation equations) can be extended and (10.9) for a finite volume release can be
in the case of relatively low particle concentra- written as
tions to consider the case of varying current
density by addition of a particle settling law 5 Fr ( g ′ p c0 Q 03 )1/ 2 c
1/ 2
(e.g., Bonnecaze et€al., 1993; Dade and Huppert, L5 / 2 = 1 − (10.10)
us
c0
1995). The most common approach is to assume
that the turbulent velocity scale within the cur- (Bonnecaze et€ al., 1993). If the excess density
rent is sufficiently high to maintain a uniform only arises from the presence of dense particles,
concentration particle suspension. However, the flow comes to a stop when c = 0, so the run-
at the base of the flow, where the current fluid out length, Lr can be found as
interacts with the underlying static boundary,
the velocities in the current decrease below that
2/5
5 Fr( g ′ p c0 Q 03 )1/ 2 (10.11)
of the settling speed of the particles, and sedi- Lr = .
mentation can occur. In this region, the particle us
concentration profile and sedimentation pro-
cess are governed by the balance of sedimenta- The particle settling speed, us, will vary with
tion and shear velocities, expressed as the Rouse particle size (and density), so equations of the
number (Table 10.1), form of (10.10) and (10.11) are applicable to
us flows containing a single particle size, whose
Pn = , (10.7)
κu settling velocity can be found from standard set-
*
tling laws such as Stokes’ Law (e.g., Dade and
where us is the settling velocity of the parti- Huppert, 1995) or Newton’s impact law (Dellino
cles, κ is the von Kármán constant (~0.4), and et€al., 2008). Extensions to polydispersed particle
u* is the shear velocity (see Valentine 1987, and suspensions have been recently proposed by
Dellino et€ al., 2008, for applications to PDCs). Harris et€al. (2002).
while the slumping dense fluid propagates particle sizes, as is typical of PDCs. Gladstone
along the channel base. This constant velocity et€ al. (1998) conducted lock-exchange experi-
regime, however, could also result from the bal- ments using bidispersed and polydispersed
ance between inertia, buoyancy, and resistance particle mixtures and found that the effects of
to the flow (Hogg and Pritchard, 2004). mixing different proportions of fine and coarse
Recent experiments have considered the ini- particles is strongly nonlinear. Adding a small
tial condition where the total depths of fluid on amount of fine particles to a current containing
each side of the lock gate are equal, but where coarse particles has a much larger influence on
the depth of the dense fluid layer in the lock flow velocity and sedimentation patterns than
is some fraction of the depth of the lock, and adding a small amount of coarse particles to a
is overlain by the same fluid as in the main current containing fine particles. Measurements
body of the tank (Gladstone et€al., 2004). In this of deposit areal density are not well reproduced
case, the density stratification in the lock leads by box models for bidispersed and polydispersed
to streamwise stratification of the resulting particle distributions (Dade and Huppert, 1995).
flow, and the stratified currents are observed Particle concentration also influences grav-
to propagate initially faster, then more slowly, ity current dynamics and deposition (Choux
than their unstratified counterparts. At early and Druitt, 2002). Lock-exchange gravity cur-
stages, gravity current propagation takes the rents containing bidisperse mixtures of dense
form of Eq. (10.6) for a fixed volume release and light particles with sizes chosen so as to be
Q0, and motion in the surrounding fluid plays in hydrodynamic equivalence produce deposits
a negligible role. Finally, as the gravity current that show normal grading of the dense particles,
energy is dissipated by displacing the surround- but the deposition of light particles depends
ing fluid and by frictional interaction with the strongly on the total particle concentration. The
underlying surface, viscous dissipation becomes light particles are deposited in hydrodynamic
important and the current enters a flow regime equivalence (i.e., at the same settling velocity)
where viscous and buoyancy forces are in bal- in dilute flows, but are segregated efficiently
ance (Huppert and Simpson, 1980). in concentrated suspensions (up to 23% by
volume; Choux and Druitt, 2002). The dynam-
10.3.2â•…Particle-laden gravity currents ics of gravity currents composed of high con-
Particle-driven gravity currents have also been centration suspensions (up to 40% by volume)
widely studied using lock-exchange experi- show an abrupt transition in deposition pattern
ments. Bonnecaze et€ al. (1993) compared the with distance from their source (Hallworth and
predictions from a box model of the form devel- Huppert, 1998), which is very different from
oped in Section 10.2 with experimental meas- the deposition profile of a lower concentration
urements of the areal density of the deposit current (e.g., Bonnecaze et€ al., 1993). Above a
from a sedimenting particle current containing critical concentration of particles, the gravity
approximately monodispersed (i.e., uniform currents stop abruptly and deposit the bulk of
size) spherical particles. They found good agree- their sediment load as a relatively thick layer of
ment with the observed current dynamics and constant thickness, with a much thinner layer
the distribution of deposit areal density with of sediment being deposited from the residual
distance from the lock. This model has also been low concentration cloud.
used to investigate the dynamics and deposition
patterns of laboratory gravity currents in which 10.3.3â•…Mixing processes
the interstitial fluid is less dense than the fluid Experiments have also been used to investigate
into which the current is propagating (Sparks more complex phenomenology of gravity cur-
et€al., 1993). rents in order to develop simplified formulations
The dynamics and sedimentation from to include in theoretical models. An important
particle-driven currents become much more process relevant to dilute PDCs is the mixing of
complex when the current contains a range of inertial gravity currents with the surrounding
fluid. The motion of turbulent billows along volume of the current, x is the distance from the
the upper surface of an inertial gravity current current source, α is a constant that describes the
which contains a low or zero concentration of amount of entrainment into the current head,
particles leads to incorporation of the surround- and ψ is a constant representing the ratio of the
ing fluid into the gravity current, or entrainment. height of the tail to the height of the head. The
Quantifying this process is important for PDCs dependence on Q1/2 indicates that the spatial
as dilution of a current by lighter ambient fluid rate of change of volume is proportional to the
reduces the density contrast and hence the height of the head. The solution to this equa-
flow velocity and run-out, even though the flow tion is
depth can increase.
1 (ψ − α ) x
2
Q (10.14)
Entrainment of fluid into a turbulent flow = 1 − ,
V0 2 V01/ 2
is difficult to calculate directly because the
flow structure is three-dimensional and time-
dependent, but entrainment can be measured in with the best fit to experimental data sug-
experiments. Hallworth et€ al. (1993) conducted gesting values α = 0.078 and ψ = 0.147 (±3%).
experiments in which an alkaline inertial fluid Application of this simplified result suggests
gravity current was released into a two-dimen- that dilution of the current can be significant
sional channel containing an acidic ambient (particle concentrations reducing from 40% vol.
fluid, with the neutralization resulting from to about 8% vol. within the first quarter of the
mixing visualized using universal indicator in total run-out distance), and that entrainment
the current. Different initial concentrations of provides an efficient mechanism for reducing
alkali in the current resulted in neutralization the particle concentration and buoyancy of
at different distances from the source, so the PDCs (Hallworth et€al., 1993). Furthermore, effi-
proportion of current and entrained fluid could cient mixing of fluid within the head of inertial
be determined with distance from the source. gravity currents and between the head and the
Entrainment was observed to take place primar- following fluid suggests that dilution will have
ily at the head of the gravity current. If the pro- an important effect throughout the depth and
portion of entrained fluid is defined as the ratio length of PDCs.
of the volume of entrained fluid to the total vol- Whereas these experimental studies pro-
ume of the current (i.e., a dimensionless quan- vide insights into the physics of dilute turbulent
tity), then dimensional considerations suggest PDCs, their potential to address dense PDCs is
that the proportion of entrained fluid depends uncertain. In this context, experimental studies
only on the initial volume of the current and on highly concentrated currents of dry granu-
distance from the source, and is independent of lar material or of gas–particle mixtures are rele-
the reduced gravity because g′ is the only quan- vant to investigate dense PDCs, as discussed in
tity including dimensions of time (Hallworth Section 10.4.
et€ al., 1993). This result was confirmed in sys-
tematic experiments, and leads to a simplified
description of entrainment by considering con-
servation of mass of the fluid in the gravity cur- 10.4╇ ╇Dynamics of granular flows
rent head,
dQ
10.4.1â•…Fundamental physics of granular
= (α − ψ )Q 1/ 2 , (10.12) flows
dx
This section describes key differences between
with the dynamics of dense granular currents propa-
gating in air, hereafter called granular flows,
Q(0) = V0,â•… (10.13) and fluid gravity currents. Granular flows most
where Q is the volume of the current head (area resemble the dense end member of PDCs, and
in a two-dimensional current), V0 is the initial are characterized by a particle volume fraction
where τ and σ are the shear and normal stress 10.4.2.â•…Experiments on dry granular
respectively. However, under steady flow con- media
ditions, the macroscopic friction coefficient,
μ(I), is shear-rate-dependent and is function of Particle interactions in steady flow
a dimensionless parameter called the inertial Particle interactions in dry, coarse-grained PDCs
number, I, which represents the ratio of the were investigated in shear-cell experiments by
microscopic timescale of particle rearrange- Cagnoli and Manga (2004, 2005). When a bed of
ment over the macroscopic timescale (1/γ̇) and pumice fragments is confined between two verti-
is also the square root of the Savage number, cal and coaxial cylinders that rest on a rough hori-
so that zontal rotating disk, energy dissipation occurs
(b)
hstop 0.2
u*
0.1
levées + channel
Flow thickness, hf
0
1 2
h*
very thin or
no deposit
rounded deposit
θ1~20° θ2~30°
Slope angle, θ (°)
within a basal collisional layer at Sa up to ~0.4. flux has decreased to zero and the flow is no
The upper layer acts as a rigid raft but moves rela- longer steady. Static borders form when down-
tive to the cylinders at a constant velocity, which slope gravitational forces are lower than fric-
is independent of the imposed shear rate, and this tional resistance at the flow margins because
suggests a frictional Coulomb behavior. In this friction is depth-dependent (cf. Eqs. 10.15 and
layer, reverse segregation of coarse light clasts and 10.16), which suggests that grain size segrega-
normal segregation of coarse dense clasts occurs. tion caused by polydispersity is not necessary
This is because the granular network expands and for levée formation (Mangeney et€al., 2007; see
the coarse and fine components have contrasting Section 10.5). The flow dynamics can be inferred
inertia when they are pushed upwards by colli- from the deposit morphology according to
sions originating at the basal layer. Pouliquen’s (1999) method, so that
Levées are common features of deposits of uf hlevee′
=B , (10.18)
coarse-grained PDCs and are reproduced in
ghlevee′ hchannel
experiments investigating steady finger-shaped
flows down rough inclines (Fig. 10.4; Félix and where uf is the flow front velocity, hlevée is the
Thomas, 2004). Levées form at the lateral static levée thickness, hchannel thickness of the deposit
borders behind the flow front, and result from in the channel, and Β is an empirical constant
emptying of the central channel once source ~0.5, typical of the granular material used by
decreasing region, resembling those in steady especially when they contain high amounts
flows down inclines (Savage and Hutter, 1989), of fine ash particles. Hence, fluidization and
and contrasting with those in shear-cell experi- related pore pressure diffusion processes can be
ments (Cagnoli and Manga, 2004). fundamental for governing their emplacement.
The flows propagate in three stages and their Fluidization can be achieved when differential
kinematics are controlled by the timescale tf = vertical motion between a (relatively ascending)
(hi/g)1/2, proportional to that of free fall of the col- gas and (relatively descending) particles is gener-
umn (Lajeunesse et€ al., 2005). The front initially ated. For PDCs, this may occur as gas is released
accelerates after release and then propagates at from the pyroclasts and/or as particles sediment
nearly constant velocity provided a > at; these two during column collapse and subsequent flow
regimes are similar to those observed in fluid grav- propagation.
ity currents (Section 10.3). The flows then enter a In the simple case of a gas flux injected at the
short stopping phase as they rapidly decelerate base of a static non-expanded granular column,
until motion ceases. There is a clear power-law as in many engineering configurations (see
dependence of the flow run-out (xf ) on a because Geldart, 1986, for review), the gas flow exerts a
the normalized mobility (xfâ•›–â•›xi)/xi is proportional drag force on the particles that increases with
to a when a < at (in both two- and three-dimen- the superficial gas velocity (ug, defined as the
sional experiments). and is proportional to a2/3 mean flow rate divided by the column cross-
(in two dimensions) or a1/2 (in three dimensions) sectional area), such that the weight of the
when a > at (Lajeunesse et€ al., 2004, 2005; Lube column can be increasingly supported and the
et€ al., 2004, 2005). Lube et€ al. (2004) concluded interparticle frictional stresses decrease. The
that the analysis is independent of any basal and dynamic pore fluid pressure, Pd, across the col-
internal friction parameter and questioned the umn increases with ug as described by the Ergun
role of Coulomb friction for most of the emplace- equation for a steady flow,
ment, until flows enter the stopping phase. In 150(1 − φ2 ) ηg u g 1.75(1 − φ)ρ g u 2g (10.20)
contrast, Balmforth and Kerswell (2005) showed Pd = + h,
φ3 d 2 φ3 d
that the normalized flow run-out depends on the
size and shape of the particles, which control the
interparticle friction. Note that an erodible sub- where h and ϕ are the height and porosity of
stratum has no influence on the flow kinematics the column, respectively, ηg the gas dynamic
and run-out (Lajeunesse et€al., 2004). In nature, the viscosity, ρg the gas density, and d the particle
polydispersity of PDCs and the presence of mod- diameter. On the right-hand side of Eq. (10.20),
erately steep slopes may complicate the processes the first and second terms are the laminar and
described above. Other experiments on granular turbulent components, respectively. The weight
column collapse have shown that the flow run-out of the particles is fully supported at ug = umf, the
is strongly dependent on the proportions of fine minimum fluidization velocity, when
and coarse particles for mixtures composed of
Pd , mf = (ρ p − ρ g )(1 − φ) gh, (10.21)
two particle sizes (Phillips et€al., 2006, Roche et€al.,
2005), and increases when an erodible substrate is
mobilized by the passage of a flow at slope angle where ρp is the density of the particles. The total
larger than ~ φr/2 for the material used (Mangeney pore fluid pressure at umf is then Pmf = Pd,mf + Ph,
et€al., 2010). where the hydrostatic component Ph = ρggh, so
that
10.4.3â•…Gas–particle flows Pmf = ρ p (1 − φ) + ρ g φ gh. (10.22)
Characteristics of gas-fluidized
pyroclastic materials Note that Eq. (10.22) simplifies to Pmf = ρp(1–ϕ)gh
For most dense PDCs the interstitial gas phase if the gas is much less dense than the parti-
may have a key influence on their dynamics, cles, as is the case for PDCs. At this stage, the
granular material is fluidized (s.s.) and has a 2007). High fines content strongly decreases umf
fluid-like behavior because stresses generated and ude, whereas temperatures above ~200°C
by inter-particle frictional contacts, which vary and shear both inhibit gas flow channeling, as
inversely with the pore fluid pressure, are neg- the former eliminates moisture-derived inter-
ligible. One obtains umf by equating the laminar particle cohesion and the latter breaks cohesive
term of the Ergun equation (first term on right- bonds and particle clusters. When these factors
hand side of Eq. (10.20)) and Eq. (10.22), so that act together, ignimbritic materials are fluidized
at umf as low as ~1 mm s−1, exhibit homogeneous
k Pmf (10.23)
u mf = , expansion, and deaerate slowly due to the high
ηg h
degree of bed expansion (up to 60–70% above
where k = ϕ3d2/[150(1–ϕ)2] is the hydraulic per- loose packing) and low ude ~ 0.5–1 cm s−1 (Druitt
meability. If the gas flux is no longer provided, et€al., 2007). The duration and degree of fluidiza-
the granular column defluidizes through a pore tion of PDCs depend on the strength and longev-
pressure diffusion process and the particle ity of gas sources but also, once gas supply has
frictional contacts rebuild. The pore pressure become ineffective, on the pore pressure diffu-
decreases with time according to sion timescale, which depends on the current
∂P ∂2 P thickness (Eq. 10.25).
(10.24)
=D 2,
∂t ∂h
The dynamics of gas–particle flows
where D = k/(ηgϕβ) is the hydraulic diffusion coef- Several experimental studies have highlighted
ficient and β is the gas compressibility, and the the fundamental role that the gas phase of
duration of pressure diffusion is proportional to PDCs can have on their emplacement dynamics
the timescale (cf. Iverson, 1997) through fluidization processes. Fluidization by
h2
air entrainment at the front of hot PDCs was pro-
(10.25)
td = . posed by McTaggart (1960) to explain increases
D
in run-out of laboratory-scale flows with tem-
For columns fluidized at ug ≥ umf, homogeneous perature, by envelopment and heating of cold
expansion (i.e., no gas bubbles) is achieved for air by hot particles, and subsequent violent
fine particles with grain sizes smaller than ~100 expansion. Air entrainment in flows of fine par-
µm (at ρp ~ 2500 kg m–3) to ~500 µm (at ρp ~ 500 ticles generated in a rotating drum is promoted
kg m–3) (Geldart, 1986), which are representative by periodic and continuous projection of clus-
of the bulk mass of most PDCs. Once defluidi- ters of particles ahead of the front (Bareschino
zation occurs, pore pressure diffusion is accom- et€al., 2008). This may cause motion-induced, self-
panied by collapse of the column at a constant �fluidization (Salatino, 2005) and thus explain the
velocity called the deaeration rate (ude), and a sedi- weakly inclined free upper surface of the flows
mentation interface that migrates upwards sep- (Fig. 10.6). This mechanism occurs because the
arates basal sedimented particles and settling defluidization timescale is much larger than the
ones above. periodicity of the avalanches.
Experimental studies of gas-fluidization of Fluidization caused by the interstitial gas
static columns of pyroclastic materials provide has also been considered in various experimen-
insights into fluidization processes of PDCs. tal configurations. The emplacement of PDCs on
Wilson’s (1980, 1984) seminal studies revealed moderate slopes (~5–25°) was investigated by
that gas flow channeling readily occurs at room Takahashi and Tsujimoto (2000), who found that
temperature due to interparticle cohesion in continuously fluidized flows of fine (d = 80 µm)
fines-rich, poorlysorted beds of ignimbrite. Later particles down inclines consist of a lower, con-
studies showed that high fines content, high tem- centrated layer (νp ~ 0.40–0.45) and an upper,
perature, and shear motion favor homogeneous thinner, and more dilute layer. In contrast to dry
fluidization with efficient support (Gravina flows, they have a nonlinear streamwise velocity
et€al., 2004; Bareschino et€al., 2007; Druitt et€al., profile whose gradient increases upwards, and a
uf
curvature that also increases with the mass flow Fr = ,
( gh f )1/ 2
rate (i.e., slope angle). Dam-break (equivalent
to lock-exchange at high g′ ~ g) experiments which are consistent with observations of iner-
on highly concentrated air–particle flows (νp ~ tial fluid gravity currents with large reduced
νp,max) were performed to investigate the propa- gravities as Fr ~ 2√2 (Roche et€ al., 2004). Their
gation and deposition mechanisms of the dense morphologies and front kinematics are the same
end member of PDCs (Roche et€al., 2004, 2008). as those of inertial water flows, with a normal-
In this configuration, columns of fine (d = 80 ized front velocity (or initial Froude number),
µm) particles are fluidized in a reservoir, and
uf (10.27)
then released into a horizontal channel which Fri = ,
1/ 2
does not contain a source of fluidizing air, so ( ghi )
that is independent of the aspect ratio of the propagate as inertial fluid gravity currents for
initial column, and which has a constant value most of their emplacement. Fluidized PDCs may
of ~√2, as commonly reported in dam-break be expanded so that their propagation is also
water flow experiments (Roche et€al., 2008). The controlled by the sedimentation rate of the par-
fluid-inertial behavior of these highly concen- ticles. In this context, experimental laboratory
trated two-phase mixtures is consistent with the investigations with pyroclastic material at high
high pore fluid pressure measured at the base temperatures reveal a structure similar to that
(Roche et€al., 2010), which may also result from of dense flows of analog material, and deaer-
significant air–particle viscous interactions (i.e., ation (with a concomitant aggrading deposit) of
high Da values, see Table 10.1). Once pore pres- the sheared mixtures occurs at a rate equivalent
sure has decreased sufficiently by diffusion, and to that determined in static beds at the same
interparticle frictional contacts rebuild, the initial expansion (Girolami et€ al., 2008, 2010).
flows enter a stopping phase and behave like Other experiments using pyroclastic material
their dry counterparts as they rapidly decelerate at high temperature were conducted by Dellino
and come to halt (Fig. 10.7). These flows consist et€al. (2007) at larger scale in order to study the
of a sliding head followed by the body, at the generation of PDCs by gravitational column col-
base of which a deposit aggrades at nearly con- lapse. Under these conditions, a basal concen-
stant rate. Both parts are sheared pervasively as trated flow and a dilute suspension of fine ash
the internal velocity increases upwards (Roche with frontal convective lobes, which are typical
et€al., 2010; see online supplements OS 10.3 and of PDCs, form readily. These large-scale experi-
OS 10.4). These results suggest that PDCs might ments offer a means of investigating the impact
have an erosive head and that their deposit is dynamic pressure of the dilute component of
built by aggradation. PDCs (Dellino et€al., 2008).
Further experimental studies have consid-
ered other characteristics of PDCs. In the dam-
break configuration, polydisperse mixtures
also propagate in a fluid-like inertial regime,
provided fine particles form a continuous net-
10.5╇ ╇Numerical modeling
work (i.e., matrix) with embedded coarser com- of PDCs
ponents that are transported passively, such
that very little segregation takes place (Fig. 10.8; 10.5.1â•… Principles
Roche et€al., 2005). This suggests that fines-rich PDCs are transported with a velocity and ther-
(i.e., matrix-supported) dense PDCs are likely to mal energy that may change with time. The
basic concept of numerical simulation is to solve subset of these equations, to the most complex
conservation equations of mass (sometimes that attempt to solve systems of several tens of
density or thickness), momentum, and energy equations, for example to describe the motions
(sometimes enthalpy). We present below a sim- of multiple particle sizes.
ple formulation of the conservation equation of Conservation equations are complex, con-
density (ρ) and volumetric momentum (ρu) of a taining partial derivatives, and, except for some
compressible flow, idealized flows, solutions can only be obtained
∂ by numerical treatment. An exact solution is
ρ + ∇ ⋅ (ρu ) = 0, (10.28) generally impossible to calculate and the stand-
∂t
1
2
ard methodology is to discretize equations in
space and often in time and obtain approxi-
∂
ρ u + ∇ ⋅ (ρuu ) = −
∇P + ∇
⋅ τ + ρ
g. (10.29)
mated solutions at the nodes of the discre-
∂t
5 6 7 tized space (i.e., Toro, 2001). Different methods
4
3
(finite differences, finite volumes, finite elem-
Equation (10.28) means that, at a given position, ents, cellular automata) and algorithms exist
the density only varies with time t (term 1), (Fig. 10.9). Their accuracy can be assessed by
according to the flux of the density that (term their ability to resolve shocks, which are dis-
2) enters or leaves this point, i.e., the product continuous variations of properties (e.g., dens-
of the spatial gradient of density (its variation ity) that can occur at the front of flows and also
in space) and the velocity at which it passes within them (e.g., sudden changes of pressure
through the position. Mass displaces but is con- at the front of explosions, and of thickness, at
served. Adding terms to the right-hand side of the front of dense flows). First-order resolution
Eq. (10.28) allows mass to be added to the sys- methods smooth the solutions of differential
tem, for example by input from a vent into a equations, meaning that shocks cannot be cal-
PDC, or mass to be removed, for example by culated correctly and changes in properties are
particle sedimentation during flow. The momen- more gradual than in reality. Consequently, gra-
tum equation (10.29) describes the conservation dients of variables at the position of the shock
of momentum per unit volume, ρu, and is a lit- are incorrect, generating strong perturbations
tle more complex because momentum is gen- in the velocity field. For simulated PDCs, this
erated and dissipated by stresses. The velocity induces large errors in velocity, thickness and
changes with time (term 3) according to the run-out, etc. Another indication of the quality
flux of momentum (term 4), but also due to the of algorithms is their numerical stability. Some
stresses that act on the flow at a given position. algorithms generate very strong unphysical
In this example, the stress is induced by pres- oscillations that disturb the overall behavior.
sure (term 5), mechanical shear stresses (term Another aspect of the numerical simulation of
6), and gravity (term 7). These stresses change PDCs is that, whatever the quality of the model,
according to the physics chosen for the simu- there is always a compromise between accur-
lated flow. To close this system of equations, acy and rapidity of calculations. The finer the
we need to complete it with initial conditions computational mesh, the more accurate the
(e.g., distribution of the density at the start of solution (as is very clear for shock restitution)
the simulation), boundary conditions (e.g., flux but also the longer the computational time
input, topography), and constitutive equations required to solve the problem.
(e.g., equations relating stress and velocity, or
equations of state relating density and tem- 10.5.2â•… Kinetic models
perature) that describe the physical behavior The earliest numerical simulations of PDCs were
of the flow. Each model formulation is a sim- made using the kinetic approach, which simu-
plified description of the natural phenomena, lates trajectories only for the flow front, and
and a wide range of numerical models exists, considers the flow as a rigid block or a mater-
from the simplest that deal only with a small ial point (Fig. 10.9(b)). Motion is calculated using
fixed vertical velocity profile depending on the inclined plane with a constant supply, and suc-
rheology chosen. cessfully reproduced the self-channeling and
The depth-averaged method has been used the frontal lobe and levée–channel morphology
to simulate granular flows in the laboratory typical of the deposits of coarse PDCs (Fig. 10.4).
(e.g., Savage and Hutter, 1989) and dense geo- They did not take into account segregation
physical flows like mud flows, landslides, induced by polydispersity, or the formation of
debris avalanches, or lahars (e.g., Heinrich levées when frictional resistance overcomes
et€ al., 2001; Pitman et€ al., 2003; Kelfoun and downslope gravitational forces at the flow mar-
Druitt, 2005; Sheridan et€ al., 2005; Lucas and gins due to the depth-dependence of friction.
Mangeney, 2007; see Chapter 14), as well as The enormous advantage of the depth-averaged
pyroclastic flows (Patra et€al., 2005; Capra et€al., approach is the speed of calculation, because the
2008; Kelfoun et€ al. 2009; Murcia et€ al., 2010). third dimension is averaged. It is thus of great
The depth-averaged method is very efficient for importance for real-time hazard assessment.
reproducing natural events if flow density can
be assumed to be constant in time and space: for 10.5.5â•… Multiphase approaches
example, it was used successfully to reproduce To account for both temporal and spatial vari-
a PDC from the 2006 eruption of Tungurahua ations of the physical properties necessary to
volcano (Kelfoun et€al., 2009; see online supple- simulate dilute PDCs, the calculation domain
ment OS 10.5). The depth-averaged approach is divided into a horizontal and vertical mesh
is more suitable for the simulation of dense (Fig. 10.9(e)). As a first approximation, it is pos-
pyroclastic flows than for dilute surges that sible to consider pyroclastic flows as a dusty gas
present strong vertical displacements at the (sometimes called a pseudogas) where particles
end of their path and strong spatial variations and gas form a homogeneous phase, move at
of density. Nevertheless, Doyle et€ al. (2010) the same velocity, and are in thermal equi-
derived a depth-averaged model in which the librium (Cordoba, 2005; Ishimine, 2005). To
two components of PDCs (dense and dilute) are develop more realistic simulations, gas and
coupled. This model reproduces the formation particle dynamics need to be distinguished
of the dense flow by sedimentation from the and treated separately. Particles are considered
dilute ash cloud. as a continuous phase, their properties being
The main problem of the present models averaged on each node in the same manner as
appears to be the rheology used. The complex the gas phase. All phases present (different size
and poorly understood physical behavior is classes of particles, gases of various composi-
generally approximated by first-order laws; a tions) share the same mesh and interact with
wide variety of rheological behavior has been each other. The first application of this approach
used, including frictional (τ = σ tanφ), constant was made by Wohletz et€al. (1984). The models
(τ = C), viscous (τ = η ∂u/∂y), Bingham (τ = C + have since become more complex as computa-
η ∂u/∂y), or Voellmy (τ = σ tanφ + ρgu2/ξ), where tional power has increased (e.g., Valentine and
y is height within the flow, η is viscosity, and ξ Wohletz, 1989; Wohletz and Valentine, 1990;
is the Voellmy coefficient. The most common Valentine et€ al., 1991; Dartevelle et€ al., 2004).
approximation uses a frictional behavior (e.g., Dobran et€al. (1993) and Neri and Dobran (1994)
Patra et€al., 2005). Assuming that particle colli- introduced one solid phase and two gas phases,
sions are important in such granular flows, Itoh air and vapor water. These codes accounted for
et€ al. (2000) used a collisional stress. Kelfoun multiple particle sizes to allow for estimation
et€ al. (2009) showed that a frictional behavior of spatial particle segregation (e.g., Neri et€ al.,
is not suitable for pyroclastic flow simulation 2003). Computational power limited those
and proposed a plastic rheology. Simulations studies to a two-dimensional (or axisymmetric)
of Mangeney et€al. (2007) considered an empir- approach and it is only recently that the first
ical variable friction coefficient (i.e., Pouliquen three-dimensional results (or four-dimensional,
and Forterre, 2002) to simulate flows over an if time is considered) were obtained (Esposti
Dade, W. B. and Huppert, H. E. (1995). A box model Esposti Ongaro, T., Cavazzoni, C., Erbacci, G., Neri,
for non-entraining, suspension-driven gravity A. and Salvetti, M. V. (2007). A parallel multiphase
surges on horizontal surfaces. Sedimentology, 42, flow code for the 3D simulation of explosive
453–471. volcanic eruptions. Parallel Computing, 33, 541–560.
Dade, W. B. and Huppert, H. E. (1996). Emplacement Félix, G. and Thomas, N. (2004). Relation between dry
of the Taupo ignimbrite by a dilute turbulent flow. granular flow regimes and morphology of deposits:
Nature, 381, 509–512. formation of levées in pyroclastic deposits. Earth
Dartevelle, S., Rose, W. I., Stix, J., Kelfoun, K. and and Planetary Science Letters, 221, 197–213.
Vallance, J. W. (2004). Numerical modeling Fisher R. V. (1966). Mechanism of deposition from
of geophysical granular flows: 2. Computer pyroclastic flow. American Journal of Science, 264,
simulations of plinian clouds and pyroclastic flows 350–363.
and surges. Geochemistry, Geophysics, Geosystems, 5, Forterre, Y. and Pouliquen, O. (2008). Flows of dense
Q08004, doi:10.1029/2003GC000637. granular media. Annual Reviews of Fluid Mechanics,
Dellino, P., Zimanowski, B., Büttner, R. et€al. (2007). 40, 1–24.
Large-scale experiments on the mechanics of Freundt, A., Carey, S. and Wilson, C. J. N. (2000).
pyroclastic flows: Design, engineering, and first Ignimbrites and block-and-ash flow deposits. In
results. Journal of Geophysical Research, 112, B04202, Encyclopedia of Volcanoes, ed. H. Sigurdsson et€al. New
doi: 10.129/2006JB004313. York: Academic Press, pp. 581–600.
Dellino, P., Mele, D., Sulpizio, R., La Volpe, L. and GDR MiDi (2004). On dense granular flows. European
Braia, G. (2008). A method for the calculation of Physical Journal E, 14, 341–365.
the impact parameters of dilute pyroclastic density Geldart, D. (1986). Gas Fluidization Technology. Wiley.
currents based on deposit particle characteristics. Girolami, L., Druitt, T. H., Roche, O. and Khrabrykh,
Journal of Geophysical Research, 113, B07206, Z. (2008). Propagation and hindered settling of
doi:10.1029/2007JB005365. laboratory ash flows. Journal of Geophysical Research,
Dobran, F., Neri, A. and Macedonio, G. (1993). 113, B02202, doi:10.1029/2007JB005074.
Numerical simulation of collapsing volcanic Girolami, L., Roche, O., Druitt, T. H. and Corpetti, T.
columns. Journal of Geophysical Research, 94, (2010). Velocity fields and depositional processes
1867–1887. in laboratory ash flows. Bulletin of Volcanology, 72,
Doyle, E. E., Hogg, A. J., Mader, H. M. and Sparks, R. 747–759.
S. J. (2010). A two-layer model for the evolution Gladstone, C., Phillips, J.C. and Sparks, R. S. J. (1998).
and propagation of dense and dilute regions of Experiments on bidisperse, constant-volume
pyroclastic currents. Journal of Volcanology and gravity currents: propagation and sediment
Geothermal Research, 190, 365–378. deposition. Sedimentology, 45, 833–843.
Druitt, T. H. (1998). Pyroclastic density currents. In Gladstone, C., Ritchie, L. J., Sparks, R. S. J.
The Physics of Explosive Volcanic Eruptions, ed. J. S. and Woods, A. W. (2004). An experimental
Gilbert and R. S. J. Sparks. Geological Society of London investigation of density-stratified inertial gravity
Special Publication, 145, pp. 145–182. currents. Sedimentology, 51, 767–789.
Druitt, T. H., Calder, E. S., Cole, P. D. et€al. (2002). Gravina, T., Lirer, L., Marzocchella, A., Petrosino, P.
Small volume, highly mobile pyroclastic flows and Salatino, P. (2004). Fluidization and attrition of
formed by rapid sedimentation from pyroclastic pyroclastic granular solids. Journal of Volcanology and
surges at Soufrière Hills Volcano, Montserrat: Geothermal Research, 138, 27–42.
An important volcanic hazard. In The Eruption of Hallworth, M. A. and Huppert, H. E. (1998). Abrupt
Soufrière Hills Volcano Montserrat, From 1995 to 1999, transitions in high-concentration, particle-driven
ed. T. H. Druitt and B. P. Kokelaar. Memoir of the gravity currents. Physics of Fluids, 10, 1083–1087.
Geological Society of London, 21, 263–279. Hallworth, M. A., Phillips, J., Huppert, H. E. and
Druitt, T. H., Avard, G., Bruni, G., Lettieri, P. and Sparks, R. S. J. (1993). Entrainment in turbulent
Maez, F. (2007). Gas retention in fine-grained gravity currents. Nature, 362, 829–831.
pyroclastic flow materials at high temperatures. Harris, T. C., Hogg, A. J. and Huppert, H. E. (2002).
Bulletin of Volcanology, 69, 881–901. Polydisperse particle-driven gravity currents.
Eames, I. and Gilbertson, M. A. (2000). Aerated Journal of Fluid Mechanics, 472, 333–372.
granular flow over a horizontal rigid surface. Heinrich, P., Boudon, G., Komorowski, J. C. et€al.
Journal of Fluid Mechanics, 424, 169–195. (2001). Numerical simulation of the December
1997 debris avalanche in Montserrat, Lesser Lucas, A. and Mangeney, A. (2007). Mobility and
Antilles. Geophysical Research Letters, 28, 2529–2532. topographic effects for large Valles Marineris
Hogg, A. J. and Pritchard, D. (2004). The effects of landslides on Mars. Geophysical Research Letters, 34,
hydraulic resistance on dam-break and other L10201, doi:10.1029/2007GL029835.
shallow inertial flows. Journal of Fluid Mechanics, Mangeney, A., Bouchut, F., Thomas, N., Vilotte, J.-P.
501, 179–212. and Bristeau, M. O. (2007). Numerical modeling
Huppert, H. E. (2006). Gravity currents: a personal of self-channeling granular flows and their
perspective. Journal of Fluid Mechanics, 554, 299–322. levée-channel deposits. Journal of Geophysical
Huppert, H. E. and Simpson, J. E. (1980). The Research, 112, F02017, doi:10.1029/2006JF000469.
slumping of gravity currents. Journal of Fluid Mangeney A., Roche, O., Hungr, O. et€al. (2010).
Mechanics, 99, 785–799. Erosion and mobility in granular collapse over
Ishimine, Y. (2005). Numerical study of pyroclastic sloping beds. Journal of Geophysical Research, 115,
surges. Journal of Volcanology and Geothermal Research, F03040, doi:10.1029/2009JF001462.
139, 33–57. Martin, D. and Nokes, R. (1988). Crystal settling in
Itoh, H., Takahama, J., Takahashi, M. and Miyamoto, a vigorously convecting magma chamber. Nature,
K. (2000). Hazard estimation of the possible 332, 534–536.
pyroclastic flow disasters using numerical McEwen, A. S. and Malin, M. C. (1989). Dynamics
simulation related to the 1994 activity at Merapi of Mount St. Helens’ 1980 pyroclastic flows,
volcano. Journal of Volcanology and Geothermal rockslide-avalanche, lahars, and blast. Journal of
Research, 100, 503–516. Volcanology and Geothermal Research, 37, 205–231.
Iverson, R. M. (1997). The physics of debris flows. McTaggart, K. C. (1960). The mobility of nuées
Reviews of Geophysics, 35, 245–296. ardentes. American Journal of Science, 258, 369–382.
Iverson, R. M. and Denlinger, R. P. (2001). Flow Mitani, N. K., Matuttis, H. G. and Kadono, T. (2004).
of variably fluidized granular masses across Density and size segregation in deposits of
three-dimensional terrain 1. Coulomb mixture pyroclastic flow. Geophysical Research Letters, 31,
theory. Journal of Geophysical Research, 106, L15606, doi:10.1029/2004GL020117.
537–552. Murcia H. F., Sheridan, M. F., Macías, J. L. and Cortés,
Kelfoun, K. and Druitt, T. H. (2005). Numerical G. P. (2010). TITAN2D simulations of pyroclastic
modeling of the emplacement of Socompa rock flows at Cerro Machín Volcano, Colombia: Hazard
avalanche, Chile. Journal of Geophysical Research, 110, implications. Journal of South American Earth Sciences,
B12202, doi:10.1029/2005JB003758. 29, 161–170.
Kelfoun, K., Samaniego, P., Palacios, P. and Barba, Neri, A. and Dobran, F. (1994). Influence of eruption
D. (2009). Testing the suitability of frictional parameters on the thermofluid dynamics of
behaviour for pyroclastic flow simulation by collapsing volcanic columns. Journal of Geophysical
comparison with a well-constrained eruption Research, 99, 11,833–11,857.
at Tungurahua volcano (Ecuador). Bulletin Neri, A., Esposti Ongaro, T., Macedonio, G. and
of Volcanology, 71, 1057–1075, doi: 10.1007/ Gidaspow, D. (2003). Multiparticle simulation
s00445–009–0286–6. of collapsing volcanic columns and pyroclastic
Lajeunesse, E., Mangeney-Castelnau, A. and Vilotte, flows. Journal of Geophysical Research, 108, 2202,
J.-P. (2004). Spreading of a granular mass on a doi:10.1029/2001JB000508.
horizontal plane. Physics of Fluids, 16, 2371–2381. Neri, A., Esposti Ongaro, T., Menconi, G. et€al. (2007).
Lajeunesse, E., Monnier, J. B. and Homsy, G. M. 4D simulation of explosive eruption dynamics at
(2005). Granular slumping on a horizontal surface. Vesuvius. Geophysical Research Letters, 34, L04309,
Physics of Fluids, 17, 103302. doi:10.1029/2006GL028597
Lube, G., Huppert, H. E., Sparks, R. S. J. and Patra, A. K., Bauer, A. C., Nichita, C. C. et€al. (2005).
Hallworth, M. A. (2004). Axisymmetric collapses of Parallel adaptive numerical simulation of
granular columns. Journal of Fluid Mechanics, 508, dry avalanches over natural terrain. Journal of
175–199. Volcanology and Geothermal Research, 139, 1–22.
Lube, G., Huppert, H. E., Sparks, R. S. J. and Freundt, Phillips, J. C., Hogg, A. J., Kerswell, R. R. and Thomas,
A. (2005). Collapses of two-dimensional granular N. H. (2006). Enhanced mobility of granular
columns. Physics Reviews E, 72, 041301, doi:10.1103/ mixtures of fine and coarse particles. Earth and
PhysRevE.72.041301. Planetary Science Letters, 246, 466–480.
Pitman, E. B., Patra, A., Bauer, A., Sheridan, M. and 1963 Little Tahoma Peak avalanches, Mount
Bursik, M. (2003). Computing debris flows and Rainier, Washington. Journal of Volcanology and
landslides. Physics of Fluids, 15, 3638–3646. Geothermal Research, 139, 89–102.
Pouliquen, O. (1999). Scaling laws in granular flows Simpson, J. E. (1997). Gravity Currents in the
down rough inclined planes. Physics of Fluids, 11, Environment and the Laboratory. Cambridge
542–548. University Press.
Pouliquen, O. and Forterre, Y. (2002). Friction law for Simpson, J. E. and Britter, R. E. (1979). The dynamics
dense granular flows: application to the motion of of the head of a gravity current advancing over a
a mass down a rough inclined plane. Journal of Fluid horizontal surface. Journal of Fluid Mechanics, 94,
Mechanics, 453, 133–151. 477–495.
Roche, O., Gilbertson, M. A., Phillips, J. C. and Sparks, Sparks, R. S. J. (1976). Grain size variations in
R. S. J. (2004). Experimental study of gas-fluidized ignimbrites and implications for the transport of
granular flows with implications for pyroclastic pyroclastic flows. Sedimentology, 23, 147–188.
flow emplacement. Journal of Geophysical Research, Sparks, R. S. J., Bonnecaze, R. T., Huppert, H. E.
109, B10201, doi:10.1029/2003JB002916. et€al. (1993). Sediment-laden gravity currents with
Roche, O., Gilbertson, M. A., Phillips, J. C. and Sparks, reversing buoyancy. Earth and Planetary Science
R. S. J. (2005). Inviscid behaviour of fines-rich Letters, 114, 243–257.
pyroclastic flows inferred from experiments on Sulpizio, R., and Dellino, P. (2008). Sedimentology,
gas-particle mixtures. Earth and Planetary Science depositional mechanisms and pulsating behaviour
Letters, 240, 401–414. of pyroclastic density currents. In Caldera Volcanism:
Roche, O., Montserrat, S., Niño, Y. and Tamburrino, Analysis, Modelling, and Response, ed. J. Gottsmann
A. (2008). Experimental observations of water-like and J. Martí, Developments in Volcanology 10.
behavior of initially fluidized, dam break granular Elsevier, pp. 57–96.
flows and their relevance for the propagation of Takahashi, T. and Tsujimoto, H. (2000). A mechanical
ash-rich pyroclastic flows. Journal of Geophysical model for Merapi-type pyroclastic flow. Journal of
Research, 113, B12203, doi:10.1029/2008JB005664. Volcanology and Geothermal Research, 98, 91–115.
Roche, O., Montserrat, S., Niño, Y. and Tamburrino, Thomas, N. (2000). Reverse and intermediate
A. (2010). Pore fluid pressure and internal segregation of large beads in dry granular media.
kinematics of gravitational laboratory air-particle Physics Reviews E, 62, 961–974.
flows: insights into the emplacement dynamics Toro, E. F. (2001). Shock-Capturing Methods for
of pyroclastic flows. Journal of Geophysical Research, Free-Surface Shallow Flows. New York: Wiley, 309 pp.
115, B09206, doi:10.1029/2009JB007133. Valentine, G. A. (1987). Stratified flow in pyroclastic
Salatino, P. (2005). Assessment of motion-induced surges. Bulletin of Volcanology, 49, 616–630.
fluidization of dense pyroclastic gravity currents. Valentine G. A. and R. V. Fisher (2000). Pyroclastic
Annals of Geophysics, 48, 843–852. surges and blasts. In Encyclopedia of Volcanoes, ed.
Saucedo, R., Macias, J. L., Sheridan, M. F., Bursik, H. Sigurdsson et€al. New York: Academic Press, pp.
M. I. and Komorowski, J. C. (2005). Modeling of 571–580.
pyroclastic flows of Colima Volcano, Mexico: Valentine, G. A. and Wohletz, K. H. (1989). Numerical
implications for hazard assessment. Journal of models of plinian eruption columns and
Volcanology and Geothermal Research, 139, 103–115. pyroclastic flows. Journal of Geophysical Research, 94,
Savage, S. B. (1984). The mechanics of rapid granular 1867–1887.
flows. Advances in Applied Mechanics, 24, 289–366. Valentine, G. A., Wohletz, K. H. and Kieffer, S. W.
Savage, S. B. and Hutter, K. (1989). The motion of (1991). Sources of unsteady column dynamics in
a finite mass of granular material down a rough pyroclastic flow eruptions. Journal of Geophysical
incline. Journal of Fluid Mechanics, 199, 177–215. Research, 96, 21,887–21,892.
Sheridan, M. F. and Malin, M. C. (1983). Application von Kármán, T. (1940). The engineer grapples
of computer-assisted mapping to volcanic hazard with nonlinear problems. Bulletin of the American
evaluation of surge eruption: Vulcano, Lipari, Mathematical Society, 46, 615–683.
and Vesuvius. Journal of Volcanology and Geothermal Wadge, G., Jackson, P., Bower, S. M., Woods, A.W.
Research, 17, 187–202. and Calder, E. (1998). Computer simulations of
Sheridan, M. F., Stinton, A. J., Patra, A. et€al. (2005). pyroclastic flows from dome collapse. Geophysical
Evaluating Titan2D mass-flow model using the Research Letters, 25, 3677–3680.
Wilson, C. J. N. (1980). The role of fluidization in the (c) Discuss the limitation of the analysis.
emplacement of pyroclastic flows: an experimental 10.2 Laboratory experiments on granular flows
approach. Journal of Volcanology and Geothermal in air (ρ0 = 1.2 kg m–3) are carried out to
Research, 8, 231–249. investigate concentrated, coarse-grained PDCs.
Wilson, C. J. N. (1984). The role of fluidization in the The particles used are glass beads of grain
emplacement of pyroclastic flows, 2: experimental size d = 2 mm and density ρp = 2500 kg m–3.
results and their interpretation. Journal of Flows are generated on an inclined rough
Volcanology and Geothermal Research, 20, 55–84. substrate and have a typical thickness hf ~
Wilson, C. J. N. (1985). The Taupo eruption New 1 cm. Calculate the range of velocities uf at
which the flows will be in a frictional or in a
Zealand, 2. The Taupo ignimbrite. Philosophical
collisional regime.
Transactions of the Royal Society of London A, 314,
10.3 An axisymmetric constant-volume gravity
229–310.
current spreads radially from its source such
Wohletz, K. H., McGetchin, T. R., Sandford, M. T.
that the position of the flow front r increases
and Jones, E. M. (1984). Hydrodynamic aspects of
with time t, from an initial condition r = 0
caldera-forming eruptions: Numerical models. at t = 0. The statement of mass conservation
Journal of Geophysical Research, 89, 8269–8285. for this current is Q = πr2hc, and the front
Wohletz, K. H. and Valentine, G. A. (1990). Computer condition when inertia and buoyancy forces
simulations of explosive volcanic eruptions. In are in balance is uc = Fr(g′hc)1/2, where Q is the
Magma Transport and Storage, ed. M.P. Ryan. London: volume and hc is the thickness of the current,
Wiley, pp. 113–135. uc is the velocity of the current front, Fr is the
Froude number, and g′ is the reduced gravity.
Using these formulae, show that the position
of the front r varies as t1/2 in the inertia-
Exercises buoyancy regime. How does the position
of the front depend on the volume of the
10.1 Consider a dilute PDC (surge) of density ρc current?
and height hc propagating on a sub-horizontal
surface with a front velocity uc in a less Online resources available at www.cambridge.
dense atmosphere of density ρ0 and height H org/fagents
(considered as semi-infinite).
(a) Calculate the Froude number at the front • Movie files of PDC simulations
of the surge. • Answers to exercises
(b) Field observations reveal that uc = 65 m s–1
and hc = 20 m.Calculate the density of the
surge (ρ0 = 1.2 kg m−3).
Magma–water interactions
Ken Wohletz, Bernd Zimanowski and Ralf Büttner
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
0.1–0.5 0.3–1.5 km
km 0–0.2
PLUME 0.1–2.0 km
km 20–50 0.5–3.0
HEIGHT 10–40
km
km km
subaqueous
hawaiian and strombolian plinian columns phreatoplinian surtseyan surtseyan lava flows
ERUPTIVE
ballistic scoria tephra fall column/collapse ash fall cypressoid jets (pillow, lineated,
PHENOMENA
tephra fall PDCs ashfall, PDCs lateral surges falls and surges folded, jumbed,
sheet)
ROLE OF
EXTERNAL Little/none Optimum Excessive
WATER
Increasing water abundance
water abundances, explosive activity is gener- or sheet lava, pillow breccia, and hyaloclastite.
ally suppressed, producing passive quenching Posteruptive hydrothermal interaction with
and thermal granulation of lava flows (but see fragmental materials may produce palagonitic
Chapter 12 for further discussion of deep-sea and zeolitic tuff, silica sinter, and travertine.
explosive activity). At somewhat more favorable Fragmental products are termed hydroclasts by
water/magma mass ratios, surtseyan activity is a Fisher and Schmincke (1984), instead of pyro-
common expression of explosive hydrovolcan- clasts, a term that refers solely to the fragmental
ism in, for example, shallow marine environ- products of explosive eruptions driven by mag-
ments. Dense, tephra-laden, cypressoid (“cock’s matic (juvenile) volatiles. Explosive hydrovol-
tail”) jets and steam clouds contain a significant canic products commonly contain significant
component of liquid water, both as a result of abundances of lithic fragments derived from
vapor condensation and direct ejection from a explosive fragmentation of country rock in vari-
shallow-water environment. Tephra is dispersed ous hydrovolcanic environments.
as fall or surges, which may be expressed as Petrographic studies of hydroclastic prod-
massive or planar-to-duneform bedded deposits. ucts involve the determination of particle size,
Phreatoplinian eruptions are the result of more shape, componentry (magmatic vs. lithic) and
intense interactions that produce a greater textural characteristics, and the chemical signa-
vapor-phase component at water–magma mass tures caused by both rapid and slow alteration.
ratios approaching the optimum for energy con- These data are indicators of the degree and type
version efficiency. Groundwater or shallow sur- of water interaction. For example, the grain size
face water is typically involved. Greater energy of hydroclasts is a function of the mass ratio of
release allows for explosive vapor expansion, water and magma involved dynamically in the
thorough magma fragmentation and the forma- interaction; grain textures are indicative of the
tion of convecting columns, and significantly type of interaction€– passive, explosive, brittle,
more widespread tephra dispersal through fall or ductile. Field characterization of hydroclastic
and pyroclastic density currents. Duneform bed- products focuses on analysis of deposit charac-
ding becomes more common for energetic surge teristics, including bedding, grading, sorting,
deposits. Transient explosive activity, i.e., vulca- lithification, and deposit thickness vs. distance
nian eruptions, may also be driven in part by from the vent. Variations of these characteristics
pressurized vapor derived from meteoric water within or among deposits can elucidate variabil-
(Chapter 7). The deposits of phreatoplinian and ity in eruptive intensity and style (e.g., fall vs.
vulcanian eruptions exhibit fewer indications of density current deposit), and degree of magma–
a liquid water phase in comparison to surtseyan water interaction.
deposits; the former are sometimes termed “dry” Experimental and field research revealed
hydrovolcanic deposits, whereas the latter are a correlation between the median grain diam-
termed “wet”. Subglacial eruptions, discussed in eters of hydroclasts and the interacting water/
detail in Chapter 13, commonly exhibit a range magma mass ratio (Fig. 11.3). In general, hydro-
of hydrovolcanic styles (from pillow effusion to volcanic tephra are distinguishable from mag-
surtseyan and phreatoplinian activity) during matic tephra by their much finer grain size.
their eruption sequence, and understanding of Furthermore, for hydrovolcanic tephra there is
magma–ice interactions benefitted greatly from an optimum water/mass ratio that produces the
recognition of tephra and deposit characteris- most thorough fragmentation; ratios less than
tics first mapped in phreatomagmatic tuff cones or greater than the optimum value result in less
and tuff rings. finely fragmented tephra.
Microscopic examination of grain shapes
11.2.4â•… Hydrovolcanic products and textures also reveals features indicative of
Hydrovolcanic solid products include tephra, hydrovolcanic origins; whereas the products of
blocks and bombs, explosion breccia, pillow purely magmatic fragmentation are dominated
(1) Initial contact and coarse mixing of fuel and 11.3.2â•… Physical properties of magma
coolant, growth of vapor film; Magma generally is treated as a three-phase sys-
(2) Quasi-coherent collapse of all vapor films in tem of melt, solids, and gases (see Chapter 4).
the premix caused by a triggering pressure The physical properties of magma are strongly
pulse (seismically induced or by local over- controlled by chemical composition, the propor-
expansion), leading to direct contact of fuel tions and types of solid content (phenocrysts,
and coolant; xenocrysts, and xenoliths), presence of exsolved
(3) Cycles of enhanced fuel–coolant heat trans- and dissolved gases, flow speed, and tempera-
fer, rapid (<1 ms) coolant expansion, fine ture. Natural silicate melts have temperatures
fragmentation of fuel, producing super- that exceed the solidus by ~100–200 K, such
heated and pressurized water, and explosive that the heat content of magma comprises a
energy release; significant proportion of latent heat. However,
(4) Volumetric expansion of fuel–coolant mix- during rapid cooling processes relevant to many
ture as superheated water transforms to magma–water interactions, the release of latent
superheated steam. heat is negligible, because quenching (and the
formation of glass) takes place, rather than
The process does not necessarily evolve through crystallization.
all these stages and may be arrested, for instance, Magma viscosity is the result of interaction
before mixing or explosion. between internal friction of the melt and
2
10 cosity depends primarily on the behavior of
dissolved volatiles, and will generally decrease
with increasing pressure. Although this pres-
sure-dependence is not yet well characterized,
it is the viscosity contrast between magma
101 and water that primarily governs interaction
10–2 10–1 100 101 dynamics, so little effect is expected on the
Shear rate, γ· (s–1) hydrodynamics of the magma–water system at
increasing pressure.
Figure 11.6╇╇ Shear stress dependent viscosity of a basaltic The intensity of magma–water interac-
melt (<5 % crystals, <2 % bubbles), at two temperatures.
tions reflects the efficiency of heat transfer
The lines are viscosity calculations from a power-law model
at the magma–water interface, which in turn
(modified from Sonder et€al., 2006).
reflects the degree of magma fragmentation.
Fragmentation can generally be viewed as tak-
mechanically coupled compressible bubbles and ing place in one of two regimes, depending upon
incompressible crystals (see Chapter 4, Section whether the characteristic deformation time is
4.5). Crystals are not only characterized by vari- greater or less than the mechanical relaxation
ous morphologies, but also by various interfacial time of the melt: (1) hydrodynamic fragmentation
couplings (e.g., wetting angles). It is therefore is restricted to deformation of two-�dimensional
not surprising that the rheology of magma needs interfacial areas (boundaries between the melt
to be addressed by a non-�Newtonian model. and gas) and is most efficient in systems sub-
Viscosity is not only temperature dependent, jected to high strain rates (rapid flow) at low
viscosities, low interfacial tension, and high A crucial parameter for the behavior of water
density contrast between the accelerating fluid at a high temperature interface is the phase
and surrounding media; (2) brittle fragmenta- transition temperature of water into steam. In
tion is the result of three-dimensional crack steady-state, equilibrium thermodynamic condi-
growth caused by strain that exceeds the elas- tions, the phase transition occurs at the boiling
tic properties of a medium (e.g., bulk modulus; point, which depends on the ambient pres-
cf. the stress and strain-rate criteria discussed sure. However, dynamic (short duration) events
in Chapter 4, Section 4.6). For hydrodynamic require application of quasi-steady-state thermo-
fragmentation to produce fine ash (< 63 μm), dynamic models. Instead of the phase change
the required accelerations are extremely high. occurring at the boiling point and spreading out-
However, brittle fragmentation can readily ward from discrete nucleation sites, the water is
produce fine ash particles through a steady heated so rapidly that it greatly overshoots its
increase in strain. The features of natural parti- boiling temperature, reaching the homogeneous
cles, experimental results, and theoretical mod- nucleation temperature (HNT), where all of the
els clearly show ash generation is dominated by water spontaneously changes state. The boiling
brittle fragmentation (Zimanowski et€al., 2003). regime of water under atmospheric conditions
The material parameter that controls the brittle- starts at 373 K and the HNT is reached at ~583
ness of magma (i.e., the mechanical deformation K. Little is known about the pressure depend-
energy needed to produce new crack surfaces) ence of the HNT. During extremely rapid heat
is the critical shear stress. The range of critical transfer, as occurs in explosive MFCI, the crit-
shear stress of magma is as large as the vari- ical point of water (22 MPa and 647 K) may be
ability of viscosity, but in contrast to viscosity, exceeded, so that water exists as a supercritical
the critical shear stress decreases with increas- fluid, with no phase boundary separating vapor
ing silica content. Therefore, basaltic magma and liquid. Experimental observations indicate
can be more than three orders of magnitude that water can remain at liquid state densities
stronger than rhyolitic magma. Bubbles and even at magmatic temperatures during such
crystals both weaken the structural strength interactions (Zimanowski et€al., 1997b).
and therefore reduce the critical shear stress. Given that > 60% of the Earth’s surface is
Further information on experimental and the- covered by oceans, magma–water interactions
oretical fragmentation studies can be found in predominantly involve seawater. Seawater is
Hermann and Roux (1990), Zimanowski et€ al. a solution dominated by the presence of salts
(1997a, 1997b, 2003), and Büttner et€ al. (1999, (mostly NaCl), and its thermodynamic behav-
2002, 2006). ior can be approximated by the two-component
system of pure water and NaCl. Figure 11.7(a)
11.3.3â•… Water physics illustrates a p–T phase diagram for the system
The thermodynamic behavior of water is well NaCl–H2O that shows phase boundaries of the
known from the steam-locomotion era, and pure components and projections of the phase
many volcanological studies have approached boundaries for intermediate compositions. The
the problem of magma–water interaction rather salinity of seawater results in critical conditions
simplistically, by direct application of the first occurring not at a single point but along a curve
law of thermodynamics. While this approach that connects the critical points of the two pure
does provide some limiting conditions, it ignores end members. These phase relationships (Fig.
complex issues concerning the multi-phase 11.7(a)) show that at any temperature two fluid
(steam, water, magmatic particles, melt) system. phases can coexist and a single critical point
Consideration of multi-phase fluid mechanics does not exist if solid NaCl is present. Critical
places further constraints on the application of behavior occurs at pressures and temperatures
thermodynamics to this problem, as described elevated from those of pure water to values
by Delaney (1982) and Wohletz (1986), and approaching ~30 MPa and ~680 K for seawater
embodied in studies of MFCI (Section 11.3.1). with a salinity of 3.2 wt % NaCl (Bischoff and
forcing, such as differential movement between (2) Metastable-film boiling: At a temperature con-
magma and water, and seismic impulses (earth- trast below the HNT, a macroscopic vapor
quakes), which can affect the stability of the film does not form; however, the thermal
magma–water interface. Experimental stud- coupling is modified by local formation of
ies suggest that external triggering is a major steam at the interface. At a temperature con-
mechanism leading to the escalation of heat trast well above the HNT, metastable-film
transfer into explosive conditions. For more boiling can take place in two cases. In Case 1,
details on magma–water mingling/mixing and high ambient pressure causes a reduction of
fragmentation physics see Morrissey et€al. (2000) the thickness of the superheated steam layer
and Zimanowski and Büttner (2002, 2003). of the film so that it reaches the same thick-
In contrast to the hydrodynamics discussed ness range as the thermal boundary layer.
in Section 11.3.3, the thermodynamic behav- In Case 2, hydrodynamic disturbances cause
ior of the magma–water interface changes interfacial instability waves with amplitudes
dramatically with increasing temperature and of the same scale as the total film thickness.
hydrostatic pressure, and the effects of min- (3) Direct contact: At a temperature contrast
eralization (e.g., seawater) are generally small below the boiling point of water and/or at
by comparison. The thermal coupling between ambient pressures exceeding the critical
magma and water is largely dependent upon point of water, the thermal coupling can be
the interface temperature and development described using equilibrium thermodynam-
of a vapor film between the magma and the ics, and it depends on the thermal proper-
water. The vapor film consists of two regions: ties of the liquids and the hydrodynamics
(1) a thicker hot layer of superheated steam at at the interface. The thermal properties of
the same temperature as and directly in con- magma are not strongly affected by increas-
tact with the magma, and (2) a thinner cool ing ambient pressure, however, the prop-
layer (the thermal boundary layer) between the erties of water change markedly (Fig. 11.8).
superheated steam layer and water, within For magma and water in direct contact, the
which condensation and vaporization balance. thermal coupling is one to two orders of
The temperature of the thermal boundary layer magnitude higher than that of stable and
can be only microscopically defined because its metastable film boiling.
thickness may be much less than a few mil-
limeters. Because the interface temperature is Considering the effect of increasing hydrostatic
a hypothetical value, dependent upon thermal pressure on the thermal coupling of magma and
diffusivities of magma and vapor, it is useful water, the stable-film boiling regime is restricted
to characterize the vapor-film by its tempera- to ambient pressures below about 1 MPa where
ture contrast with the magma (i.e., the magma the vapor film thickness at the magmatic tem-
surface temperature minus water tempera- peratures of the hot layer becomes critical.
ture). In so doing, the transfer of heat energy Above ambient pressures of 10 MPa, the moder-
between substances, known as thermal coupling, ating effect of metastable-film boiling can prac-
can be classified in three regimes, with coup- tically be excluded. At thermal power plants,
ling increasing from (1) to (3): where an optimum heat flux (i.e., direct con-
tact) is required, the water pressure of the heat
(1) Stable-film boiling: At a temperature contrast exchange system usually is set to 15–20 MPa for
well above the water HNT and an ambient temperature contrasts comparable to magmatic
pressure well below the critical pressure of conditions. The thermal coupling at a constant
water, a macroscopic vapor film forms at the magma–water interface at water depths exceed-
interface between magma and water and ing ~1 km (equivalent to 10 MPa) can therefore
restricts the thermal coupling. The thickness be described well by the direct contact regime.
of the vapor film is greater for high tempera- Magma fragmentation is an important
ture contrasts and low ambient pressures. aspect of the interface physics. Wohletz (1983)
discusses a number of hydrovolcanic fragmen- material relative to the shock front, and ur must
tation mechanisms that create a complex inter- equal the sonic velocity of the shocked mater-
face geometry, but the actual fragmentation ial (Courant and Friedrichs, 1948; and Zel’dovich
mechanisms involved are generally categorized and Raizer, 1966).
as hydrodynamic (ductile) and brittle (Section The C–J condition can be evaluated on a pres-
11.3.2). With increasing fragmentation, the sure–volume diagram (Fig. 11.9) that shows the
interfacial surface area increases exponentially. shock adiabat (termed the Hugoniot and defined
This increase in surface area is often referred to as the locus of points representing pressure–
as mixing. The resulting mixture evolves from volume states achievable by shocking a mater-
a pre-expanded combination of magma frag- ial from an initial state) and the release adiabat
ments and high-pressure water and vapor to a (called the Rankine–Hugoniot curve or detonation
post-expanded mixture of quenched fragments adiabatic). These adiabats are concave upward
and steam, which typically forms an eruption and the detonation adiabatic exists at higher
column. Although Kokelaar (1986) distinguishes volume states than the Hugoniot. The deton-
contact-surface interaction (dynamics along an inter- ation adiabatic is defined by classical Rankine–
face between a free body of water and magma) Hugoniot jump conditions for conserving mass,
from bulk interaction (dynamics of a volume of momentum, and energy across a shock wave
magma that confines water or water-rich clas- (Landau and Lifshitz, 1959; Zel’dovich and
tic materials), both cases involve heat transfer Kompaneets, 1960; and Zel’dovich and Raizer,
along the interface between magma and water. 1966). These conditions can be stated as:
Experimental evidence indicates that most of
ρ1u1 = ρ2 u2 ; (mass) (11.1)
the fragmentation occurs during development
of interface dynamics and prior to expansion p1 + ρ1u12 = p2 + ρ2 u22 ; (momentum)
(11.2)
and eruption. Nevertheless, Mastin (2007) devel-
ops a strong argument for the occurrence of tur- 1
bulent shedding of glassy rinds from fragments
E2 − E1 = ( p1 + p2 ) ( V1 − V2 ) ; (energy ) (11.3)
2
in erupted jets of steam and ash. The physics of
the dynamic interface are complex and poorly
understood; however, experimental studies of
shock waves associated with interface dynam-
ics suggest the physics may involve detonation, Von Neuman Rankine–Hugoniot Curve
Point (detonation adiabatic)
i.e., the formation of an exothermic shock front
whose propagation creates an explosion.
Wohletz (2003) discusses how film boiling
Pressure
Hugoniot
with metastable film boiling, it is also hydrody-
e
(shock adiabatic)
namically unstable because of large pressure,
density, sound speed, and conductivity gra-
(p1, V1)
dients produced by the film. Such instability
is prone to a kind of detonation, termed ther-
mal detonation, especially if perturbed by some
Volume
external pressure wave, such as that produced
by volcanic seismicity. In a thermal detonation, Figure 11.9╇╇ Pressure–volume diagram showing the
the acceleration of the particle mixture by the relationships among the initial state of the magma–water
shock wave must produce a relative velocity, ur, mixture (p1, V1), the shock Hugoniot, the Rayleigh line, the
high enough to satisfy the Chapman–Jouguet Von Neumann and Chapman–Jouguet (C-J) points, and the
(C–J) condition: ur is the speed of the shocked detonation adiabatic.
where ρ is mixture density, V is mixture spe- their MFCI experiments. In order to assess the
cific volume = 1/ρ, p is pressure, u is mixture role of ambient pressure in hydrovolcanism
velocity, E = E(p,V) is mixture internal energy, using detonation theory, the method of Board
and subscripts 1 and 2 indicate the unshocked et€ al. (1975) can be used to calculate (using
and shocked states. All notation is given in the Rankine–Hugoniot jump conditions) the
Section 11.6. For an inert material, the shocked speed of a propagating shock wave through a
thermodynamic states given by Eq. (11.3) define magma–water mixture. For mixture condi-
the Hugoniot, which is typically determined tions of R ~ 0.5 (where R is the mass ratio of
experimentally to derive the equation of state water to magma), a shock speed is ~300 m s−1.
of a material. Two points on the Hugoniot, one This velocity defines a Rayleigh line slope, and
at the initial pressure (p1,V1 in Fig. 11.9) and the then the Rankine-Hugoniot conditions must be
other at the pressure of the shock front (the specified in order to predict a detonation adia-
von Neumann point or spike, since there is a tran- batic that touches the Rayleigh line at a sin-
sient pressure peak at this location), define the gle point of tangency, the C–J condition of (pcj,
Rayleigh line. For a reactive material in which Vcj). To induce mechanisms for melt breakup
the shock induces vapor production in its wake (Fig. 11.10), the C-J condition requires that the
(either chemically for common explosives, or slip velocity us between the shocked melt frag-
physically for a magma–water mixture where ments and water is at least as large as ur. Using
nearly instantaneous heat transfer produces the classical detonation theory described by
vapor), the energy equation (Eq. 11.3) gives Eqs. (11.1)–(11.3), the relative velocity of the
all possible states of the detonation adiabatic shocked mixture leaving the front is a function
(Fig.€11.9). of the mixture pressure and specific volume at
By combining Eqs. (11.1) and (11.2), one initial and C–J conditions:
obtains the value of (p2€ – p1)/(V1€ – V2), which is
a constant defined by the slope of the Rayleigh u r = ( pcj − p1 )(V1 − Vcj ). (11.4)
line. The square root of this slope is propor-
tional to the velocity of the detonation wave. For an idealized thermal detonation in which
Zel’dovich and Kompaneets (1960) show that the pcj ≈ 100 MPa (Board et€al., 1975), Eq. (11.4) sets ur
only possible steady state (in which a detonation at ~100 m s−1. For volcanic MFCIs, the approach
wave is sustained) is where the Rayleigh line is of Corradini (1981) yields a minimum ur of 60
a tangent to the detonation adiabat at the C–J m s−1 (Wohletz, 1986). Drumheller (1979) com-
point. The points behind a propagating shock at bined the requirements for relative velocity and
which the C–J condition exists define a surface melt breakup time into a critical Bond number (the
known as the C–J plane or the detonation front (not Bond number is ratio of body and surface ten-
to be confused with the shock front). sion forces). By assuming a constant pcj, Wohletz
Figure 11.10 shows a generalized concep- (1986) evaluated the critical Bond number with
tual view of thermal detonation that includes respect to MFCI experimental data (Wohletz
observed phenomena such as thermohydrau- and McQueen, 1984) to predict the effects of R
lic fracturing and brittle reaction, described and ambient pressure on the development of
by Zimanowski et€ al. (1997b) and Büttner and relative velocities and magma particle sizes.
Zimanowski (1998), which lead to enhanced Optimal conditions for thermal detonation
heat transfer and catastrophic vapor expan- exist at 0.5 < R < 2.0, for ambient pressures
sion. Zimanowski et€ al. (1997a) document ≤ 40 MPa (Wohletz, 2003). With increasing ambi-
experiments that show development of intense ent pressure, predicted relative velocities fall to
shock waves in less than a millisecond under < 60 m s−1 (which is considered the lower limit
extreme rates of cooling (> 106 K s−1) and stress for sustaining a detonation; Wohletz, 1986), and
(> 3 GPa m−2). Yuen and Theofanous (1994) particle fragmentation decreases, meaning less
demonstrate that application of detonation thermal energy is released in the wake of the
theory successfully predicts results of many of shock wave.
measured directly. To verify the relevance to direct contact producing about two orders of
natural volcanic processes, the experiments magnitude greater heat flux per unit area than
were designed to produce artificial tephra in stable film boiling (Section 11.3.4). Three heat
statistically relevant quantities. Furthermore, flux regimes in the magma–water system can be
the products were nearly identical to the nat- defined to allow classification of magma–water
ural analogs in terms of grain-size, morphology, interaction and facilitate theoretical descrip-
and chemical composition (Fig. 11.12). tion of the thermodynamics and hydrodynam-
Experimental results demonstrate that the ics of the system (Wohletz, 1995; Büttner and
intensity of magma–water interaction, and thus Zimanowski, 1998; Zimanowski et€ al., 2003;
the danger potential of hydrovolcanic erup- Büttner et€al., 2005):
tions, depends on the efficacy of the heat trans-
(1) Non-explosive interaction regime. The heat flux
fer from magma to water, i.e., the amount of
in the system does not create a water over-
heat transferred per unit volume and time. As
pressure (subcritical state of coolant). Steady
discussed in Section 11.3.4, this heat transfer is
thermodynamic sink conditions are main-
directly correlated to the interfacial area (i.e.,
tained (i.e., water can always be described
the size of the contact area between magma and
as a passive heat sink and steady-state
water per unit volume) and the interfacial coup-
thermodynamic models are applicable).
ling conditions. Heat transport in the magma to
Fragmentation of magma is governed by its
the interface is controlled by the temperature-
rheology in an aqueous environment (e.g.,
dependent magma thermal conductivity; con-
pillow formation) and/or by thermal con-
vection can be neglected because of the high
traction of melt (thermal granulation), lead-
viscosity of magma and the short timescale.
ing to passive fragmentation (Schmid et€al.,
Heat transport in the water (having a viscos-
2010; Sonder et€al., 2011).
ity at least three orders of magnitude less than
(2) Subsonic explosive interaction regime. The heat
magma) away from the interface is controlled by
flux in the system creates a water overpres-
both conduction and convection, and therefore
sure (critical state of coolant). Superheated
is in principle much more effective. The ther-
water is generated and complex phase tran-
mal coupling at the interface can be described
sitions occur. Steady-state thermodynamic
in the three regimes: of stable film boiling,
models need significant modifications, but
metastable film boiling, and direct contact, with
still are generally applicable. Fragmentation
(of magma is dominated by brittle proc- (e.g., water vaporization and magma crystalliza-
esses caused by the hydraulic forcing of the tion). During explosive interaction the energy
coolant, but governed by the mechanical exchange happens rapidly, resulting in a min-
properties of the magma (i.e., mechanical imum of energy going towards phase transitions
deformation does not exceed critical shear and most going towards raising the temperature
conditions and propagation of cracks is sub- of the water and cooling the magma. For this
sonic with respect to the shear-wave vel- simple situation, the equilibrium temperature,
ocity). This regime leads to subsonic active Te, is between the temperature of the magma,
fragmentationAustin-Erickson et€al., 2008). Tm, and that of the water, Tw. Therefore, simple
(3) Supersonic explosive interaction regime. The heat energy conservation can be expressed as:
flux in the system increases rapidly, due to
a thermohydraulic feedback mechanism m w c w (Te − Tw ) = m m c m (Tm − Te ) , (11.5)
and high overpressure is generated in the
water (supercritical state of coolant). Non- where m is the mass and c is the specific heat
equilibrium thermodynamic models are capacity of the water (subscript w) and the
recommended. Fragmentation of magma magma (subscript m), respectively. Equation
is driven by hydraulic forcing of the cool- (11.5) shows that Te varies with the mass ratio
ant exceeding the mechanical properties of water to magma, mw/mm, denoted as R. In add-
and propagation of cracks is supersonic. ition, the specific heat ratio of water to magma
Consequently a significant proportion (up is typically in the range of 3.0 to 4.0, depending
to 80%) of the mechanical energy is released upon magma composition and the range of typ-
as shock waves. This regime leads to super- ical water states at the surface of the Earth. For
sonic active fragmentation. most cases, it is assumed that the ratio ξ = cw/cm
~ 3.5. Accordingly, we can rearrange Eq. (11.5)
The consequences for the eruptive behavior of to predict Te:
the magma–water system depend on the inter- ξ RTw + Tm
acting magma/coolant volume ratio, the rheo- Te = . (11.6)
1 + ξR
logical properties of the magma, the thermal
and rheological properties of the coolant, the The value of Te can be thought of as an idealized
hydrodynamic mingling energy, the ambient initial thermal equilibrium prior to explosive
pressure, and the geometry of the mingling vaporization of the water, which allows thermo-
space. In the following section, we discuss dynamic prediction of the resulting mechan-
modeling techniques for steady state regimes ical energy of the interaction. Figure 11.13(a)
1 and 2 and the non-equilibrium conditions of shows some predictions of Te as a function of
regime 3. R. However, this thermal equilibrium is just an
idealization and many factors can cause initial
11.4.2â•…Multiphase equilibrium interaction temperatures to be higher or lower,
thermodynamic models one of which is the composition of the water.
Simple conservation of energy provides for a Water at the surface of the Earth may con-
first-order assessment of magma–water interac- tain dissolved or suspended constituents (e.g.,
tions. For most cases, the magma temperature muddy water), or it may be contained within
exceeds the water vaporization temperature, pores of rock or sediments. The effect on the
which varies somewhat with composition thermal equilibrium can be evaluated as follows.
and ambient pressure. During interaction, the For interactions between magma and saturated
internal energy of the water increases by an rocks or impure water (White, 1996), rock and
amount equal to that lost by the magma. While impurities act as heat sinks, and the mass ratio,
most of that internal energy is involved in heat- Rr, is given as
ing the water and quenching the magma, some mr
Rr = , (11.7)
of it may also be involved in phase transitions mm
Sw φ ρw
xw = , (11.9)
ρr
rock, allowing gradual quenching. With grad- In addition, the rapid heat loss from the magma
ual quenching, the film slowly heats water in by this continued vapor-film instability leads to
the rocks near the magma at a rate balanced magma quenching and possible granulation.
by the heat transfer away from it by the con- With these interface phenomena, the magma
vective movement of pore water, provided that becomes increasingly fragmented, leading to
pores are interconnected. However, such passive larger surface areas for heat transfer and larger
quenching is not always the case. Consider the volumes of superheated water and vapor.
hypothetical instantaneous interface tempera- Further thermodynamic predictions about
ture, TI, attained by the initial contact of magma magma–water interactions can be derived from
with water that can be estimated (Cronenburg, analysis of the thermodynamic work done by
1980) by: the expansion of water from its initial thermal
equilibrium. We have assumed that nothing
TI =
(
Tm k m ) (
κ m + Tw k w κw ), (11.12)
is known about the mechanism of the contact
(k m κm ) + (k
w κw ) between magma and water, but that it results in
production of high-temperature and high-pres-
where k and κ are thermal conductivity and sure water that may explosively decompress.
diffusivity, respectively, and subscripts m and Thermodynamic work is manifested in the frac-
w refer to magma and water, respectively. For ture and excavation of country rock to form a
the contact of a typical basalt magma with pure crater, fragmentation of the magma into fine-
water, TI approaches 1000 K. Because of the rap- grained debris, and ejection of these fragments
idity of heat exchange, water may exist in the in an expanding jet of steam. For a hydrovol-
metastable state of superheating in which it is a canic eruption, it is necessary to find the work
liquid well above its vaporization temperature. potential for expansion of the steam to atmos-
A consequence of this superheated state is that pheric pressure. Calculation of the true poten-
it continues to absorb heat at a high rate, reach- tial requires determination of a complex set of
ing temperatures well in excess of its spontan- boundary conditions unique to each eruption,
eous nucleation temperature (i.e., HNT ~583 but for simplicity and generality, one can make
K). As temperatures approach the critical tem- some standard assumptions that allow analyt-
perature (647 K), instantaneous vaporization by ical calculation of maximum potential. For the
homogeneous nucleation produces a vapor film system consisting of a mixture of magmatic par-
that expands rapidly and is highly unstable€ – ticles and water and the surroundings being the
that is, it can expand well beyond the thermo- volcano vent structure and the atmosphere, we
dynamic equilibrium thickness. In so doing, the assume that:
vapor becomes supercooled, leading to spon-
(1) All heat lost by the magma during the erup-
taneous condensation. The condensation then
tion is transferred to external water;
leads to a rapid collapse such that liquid water
(2) Liquid water and magma are incompressible,
impacts the magma surface with a finite amount
so that for each the specific heats remain con-
of kinetic energy, leading to a second spontan-
stant with changing pressure and volume;
eous vaporization event. This cyclic vapor film
(3) The specific volume (reciprocal of density) of
growth and collapse is repeated continuously,
liquid water is small compared to that of its
typically with a frequency of up to 1 kHz (analo-
vapor;
gous to the Leidenfrost phenomenon of a drop of
(4) Water vapor behaves as a perfect gas and the
water vibrating on a hot metal surface). Vapor
magma volume does not change during the
film instability can generate enough kinetic
eruption (i.e., it does not vesiculate or con-
energy to distort the interface between the
tract during cooling).
magma and wet rock, as well as cause failure of
the host rock. In some cases, film collapse can Some heat is lost from the magma to both the
lead to jets of water-saturated rock fragments country rock and the atmosphere during erup-
penetrating the magma surface (White, 1996). tion, and magma does exhibit a finite volume
11.4.3â•…A non-equilibrium
thermodynamic model
In the supersonic explosive interaction regime
(regime 3; Section 11.4.1), water reacts as a super-
critical fluid that undergoes complex phase tran-
sitions that are difficult to model. Consequently,
the steady-state models described in Section
11.4.2 do not fully capture some important
physics. Experimental observations reveal that
magma quenches with little or no phase change
and can therefore be described as a supercooled
liquid (i.e., glass; Section 11.3.2), the physical
properties of which are well known and/or can
easily be measured. It is therefore much simpler
to model the non-equilibrium behavior of the
magma rather than the complex phase reac-
tions of the water.
The basis of the model is the heat conduc-
tion equation of standard, non-equilibrium ther-
modynamics applied to the individual magma
fragments:
∂T( x , y , z ,t )
= k m ∇2T( x , y , z , t ) , (11.13)
∂t
Ebert, H.-P., Hernberger, F., Fricke, J. et€al. (2003). Jaggar, T. A. (1949). Steam blast volcanic eruptions.
Thermophysical properties of a volcanic rock Hawaii Volcano Observatory, 4th Special Report.
material. High Temperatures€– High Pressures, 34, Kieffer, S. W. and Delany, J. M. (1979). Isentropic
561–668. decompression of fluids from crustal and mantle
Fisher, R. V. and Schmincke, H.-U. (1984). Pyroclastic pressures. Journal of Geophysical Research, 84,
Rocks. Berlin: Springer-Verlag, 472 pp. 1611–1620.
Fisher, R. V. and Waters, A. C. (1970). Base-surge bed Kokelaar, P. (1986). Magma-water interactions in
forms in maar volcanoes. American Journal of Science, subaqueous and emergent basaltic volcanism.
268, 157–180. Bulletin of Volcanology, 48, 275–290.
Frazzetta, G., La Volpe, L., and Sheridan, M. F. (1983). Krauskopf, K. B. (1967). Introduction to Geochemistry.
Evolution of the Fossa cone, Vulcano. Journal of New York: McGraw-Hill.
Volcanology and Geothermal Research, 17, 329–360. Landau, L. D. and Lifshitz, B. M. (1959). Fluid
Fuller, R. E. (1931). The aqueous chilling of basaltic Mechanics: Volume 6, Course of Theoretical Physics. New
lava on the Columbia River Plateau. American York: Pergamon.
Journal of Science, 21, 281–300. Mastin, L. G. (2007). Generation of fine
Grady, D. E. (1982). Local inertial effects in dynamic hydromagmatic ash by growth and disintegration
fragmentation. Journal of Applied Physics, 53, of glassy rinds. Journal of Geophysical Research, 112,
322–523. B02203, doi:10.1029/2005JB003883.
Grady, D. E. (1985). Fragmentation under impulsive Moore, J. G., Nakamura, K. and Alcaraz, A. (1966).
stress loading. In Fragmentation by Blasting, ed. W. L. The 1965 eruption of Taal Volcano. Science, 151,
Fourney and R. R. Boade. Society for Experimental 995–960.
Mechanics, Brookfield Center, 63–72. Morrissey, M., Zimanowski, B., Wohletz, K.,
Greer, S. C. and Moldover, M. R. (1981). and Büttner, R. (2000). Phreatomagmatic
Thermodynamic anomalies at critical points Fragmentation. In Encyclopedia of Volcanism, ed. H.
of fluids. Annual Reviews of Physical Chemistry, 32, Sigurdsson. London: Academic Press, pp. 431–445.
233–265. Muffler, L. J. P., White, D. E. and Truesdell, A.
Heiken, G. (1971). Tuff rings: examples from the Fort H. (1971). Hydrothermal explosion craters in
Rock-Christmas Lake Valley, south-central Oregon. Yellowstone National Park. Geological Society of
Journal of Geophysical Research, 76, 5615–5626. America Bulletin, 82, 723–740.
Heiken, G. and Wohletz, K. H. (1985). Volcanic Ash. Nairn, I. A. and Solia, W. (1980). Late Quaternary
Berkely, CA: University of California Press, 245 hydrothermal explosion breccias at Kawerau
pp. geothermal field, New Zealand. Bulletin of
Heiken, G. and Wohletz, K. (1991). Fragmentation Volcanology, 43, 1–13.
processes in explosive volcanic eruptions. In Noe-Nygaard, A. (1940). Subglacial volcanic activity
Sedimentation in Volcanic Settings, ed. R. V. Fisher and in ancient and recent times. Folia Geograph. Danica,
G. Smith, Society of Sedimentary Geology Special 1, no. 2.
Publication 45, 19–26. Ollier, C. D. (1974). Phreatic eruptions and maars.
Henry, R. E. and Fauske, H. K. (1981). Required initial In Physical Volcanology, ed. L. Civett, P. Gasparini,
conditions for energetic steam explosions. In Fuel– G. Luongo, and A. Rapolla. New York: Elsevier,
Coolant Interactions. New York: American Society of 289–310.
Mechanical Engineers, Report HTD-V19. Ruggles, A E., Drew, D. A., Lahey, R. T., Jr. and
Herrmann, H. J. and Roux, S. (1990). Statistical Models Scarton, H. A. (1989). The relationship between
for the Fracture of Disordered Media, Amsterdam: standing waves, pressure pulse propagation and
North Holland. the critical flow rate in two-phase mixtures. Journal
Holloway, J. R. (1977). Fugacity and activity of of Heat Transfer, 111, 467–473.
molecular species in supercritical fluids. In Ruggles, A. E., Vasiliev, A. D., Brown, N. W. and
Thermodynamics in Geology, ed. D. G, Fraser. Wendel, M. W. (1997). Role of heater thermal
Dordrecht, Holland: Reidel, pp. 161–181. response in reactor thermal limits during
Honnorez, H. and Kirst, P. (1975). Submarine oscillatory two-phase flows. Nuclear Science and
basaltic volcanism: morphometric parameters Engineering, 125, 75–83.
for discriminating hyaloclastites from hyalotuffs. Schmid, A., Sonder, I., Seegelken, R. et€al. (2010).
Bulletin of Volcanology, 39, 441–465. Experiments on the heat discharge at the dynamic
Zel’dovich, Ya. B. and Kompaneets A. S. (1960). Theory explosions. Journal of Volcanology and Geothermal
of Detonation. London: Academic Press. Research, 48, 341–358.
Zel’dovich, Ya. B. and Raizer, Yu. P. (1966). Physics Zimanowski, B., Büttner, R., Lorenz, V. and H.-G.,
of Shock Waves and High-Temperature Hydrodynamic Häfele (1997a). Fragmentation of basaltic melt
Phenomena, Volumes I and II. New York: Academic in the course of explosive volcanism. Journal of
Press. Geophysical Research, 107, 803–814.
Zimanowski, B. (1998). Phreatomagmatic explosions. Zimanowski, B., Büttner, R. and Nestler, J. (1997b).
In From Magma to Tephra, ed. A. Freundt and M. Brittle reaction of a high temperature ion melt.
Rosi. Developments in Volcanology 4, Amsterdam: Europhysics Letters, 38, 285–289.
Elsevier, pp. 25–54. Zimanowski, B., Wohletz, K. H., Büttner, R. and
Zimanowski, B. and Büttner, R. (2002). Dynamic Dellino, P. (2003). The volcanic ash problem. Journal
mingling of magma and liquefied sediments. In of Volcanology and Geothermal Research, 122, 1–5.
Peperite, ed. I. Skilling, J. D. L. White and J. McPhie.
Amsterdam: Elsevier, pp. 37–44. Online resources available at www.cambridge.
Zimanowski, B. and Büttner, R. (2003). org/fagents
Phreatomagmatic explosions in subaqueous
eruptions. In Explosive Subaqueous Volcanism, ed. J. • Supplement 11A: Calculation of interaction thermo-
D. L. White, J. L. Smellie and D. Clague. American dynamic work
Geophysical Union Monograph 140. Washington, • Supplement 11B: Video of MFCI experiments
pp. 51–60. • Links to websites of interest
Zimanowski, B., Fröhlich, G. and Lorenz, V. (1991).
Quantitative experiments on phreatomagmatic
Deep-sea eruptions
Tracy K. P. Gregg
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
multiple seamount eruptions for the first time, the oceanic lithosphere by a larger volcanic
associated with seamounts in the submarine pile (such as those around the Hawaiian arch;
Pacific “Ring of Fire,” including West Mata vol- Torresan et€ al., 1991; Bridges, 1997 (http://wal-
cano (15.10°S, 173.75°E, with a summit ~1200€m rus.wr.usgs.gov/gloria/hwgloria/)); or (4) linear
below sea level) (Baker et€ al., 2010; Chadwick seamount chains associated with, and oriented
et€ al., in review) and Northwest Rota-1 volcano roughly perpendicular to, mid-ocean ridges (e.g.,
(14.60°S, 155.78°E with a summit ~500 m below the Lamont seamount chain near 10°N off the
sea level; Embley et€ al., 2006, 2007; Chadwick East Pacific Rise (Allan et€al., 1989).
et€al., 2008a). Although seamounts associated with intra-
Chapter 11 discusses magma–water inter- plate activity and with mid-ocean ridges tend
action in shallow crustal or surface environ- to be basaltic (Batiza, 1982), seamounts formed
ments, and the role that external water plays in subduction zones and in back-arc basins dis-
in producing vigorously explosive subaerial play a range of eruptive products and compo-
eruptions. This chapter examines eruptions sitions, including, but not limited to, tholeiite
under great depths of water. The tectonic set- basalt, alkalic basalt, boninite (Embley et€ al.,
tings in which submarine volcanism occurs are 2007; Tamura et€ al., 2010), and lavas contain-
first presented, followed by a discussion of how ing as much as 70 wt.% SiO2 (Stern et€al., 2010).
the ambient conditions of the deep ocean floor This range of compositions can affect eruption
affect submarine eruptive processes. Inferences behavior and the volcanic products (Deardorff
gleaned from deep-sea eruptions can provide et€al., 2011).
useful insights into eruptive processes in other Seamounts are perhaps most familiarly
high-pressure environments, such as on Venus found as part of seamount chains such as the
(Chapter 17), which has a surface atmospheric Hawaiian-Emperor seamount chain that extends
pressure equivalent to the pressure beneath west and northwest of the Big Island of Hawaii.
~€900 m of water. Wilson (1963) proposed that seamount chains
are formed by the passage of Earth’s litho-
spheric plates over a stationary mantle plume,
or hot spot. Although there is not complete con-
12.2╇ ╇Settings for submarine sensus in the scientific community about the
volcanic deposits existence or behavior of these mantle plumes,
their presence most simply explains those sea-
12.2.1â•… Volcanic seamounts mount chains that display a well-constrained
Seamounts are submarine volcanic mountains age progression. Linear to arcuate seamount
taller than ~100 m. No one knows the precise chains without a clear age progression (e.g., the
number of seamounts in Earth’s oceans because Wolf-Darwin lineament, located on the southern
most of the ocean floor has yet to be surveyed flank of the Galapagos spreading center) require
at sufficiently high resolution, but there could other mechanisms (Ito et€al., 2003).
be as many as 25 million of them (Wessel et€al., Seamounts clustered around mid-ocean
2010, and references therein). Thus, not all ridges tend to have compositions similar to
deep submarine volcanism is constrained to MOR lavas. Fast-spreading centers display more
mid-ocean ridges. Seamounts can be classified off-axis seamounts than do slow-spreading cent-
as belonging to one of the following groups: ers (Fig. 12.1; see online supplement OS12A for
(1) part of a linear or arcuate seamount chain color versions of figures). They tend to be small
(the Hawaiian-Emperor seamount chain being in volume (< 3 km3), and are likely the result
among the best known) that are related to intra- of “leaky” MOR magma chambers (e.g., Fornari
plate mantle plumes or subduction (such as et€al., 1987; Perfit and Chadwick, 1998).
the Marianas volcanic arc); (2) associated with Seamount fields are associated with exten-
mid-ocean ridges (e.g., Smith and Cann, 1992; sional flexure of the oceanic lithosphere, both
Cochran, 2008); (3) associated with flexure of around regions of volcanic loading (such as
around the Hawaiian arch) and subduction zones the total crustal thickness of the Ontong-Java
(such as in the northwest Pacific; Valentine and Plateau is between 30 and 43 km, giving a total
Hirano, 2010). Although research into these volume of > 5 × 107 km3 (Neal et€al., 1997, and
types of seamounts is limited, they may hold references therein).
some similarities to subaerial monogenetic bas- An active eruption of this type has not yet
alt fields (Valentine and Hirano, 2010). been observed€ – on sea or on land€ – so erup-
Note that the formation and evolution of tion processes must be inferred. Drilling into
seamounts is a rich area of research; please see the Ontong-Java plateau reveals that the plat-
the online supplementary material (listed at end eau is composed of extrusive lava flows, with
of chapter) for further reading on this topic. no record of explosive deposits: roughly half of
the returned cores were composed of pillowed
12.2.2â•…Oceanic plateaus and large flows, and half of massive (probably inflated
igneous provinces sheet) flows (Neal et€ al., 1997). Given that the
Large igneous provinces (LIPs) are accumula- lavas were emplaced on a relatively level sur-
tions of mafic (typically basaltic) lavas that: face, pillowed lavas suggest a lower effusion
(1) are emplaced in a short period of time rate than both sheet lavas (Fink and Griffiths,
(a few million years); (2) cover large areas 1992; Gregg and Fink, 1995) and the large-scale
(>105 km2); and (3) comprise 105 to 106 km3 inflated pahoehoe flows identified in subaerial
(Coffin and Eldholm, 1994). Submarine LIPs LIPs (Self et€al., 1998). Similar to subaerial LIPs,
have been studied primarily by drilling (Ocean vents and vent regions have not been identi-
Drilling Program, 1991, 2001) and through ship- fied in the submarine oceanic plateaus. Coffin
board sonar mapping (Neal et€al., 1997). The sub- and Eldholm (1994) identified 34 oceanic plat-
marine Ontong-Java Plateau is an Alaska-sized eaus and oceanic flood basalt provinces, each
feature centered around 2.08°S, 158.09°E, and of which is composed of countless individual
is one of the largest LIPs yet identified (Coffin submarine lava flows. The lack of hyaloclastites
and Eldholm, 1994; Neal et€ al., 1997). As the within these oceanic plateaus suggests that nei-
descriptor “plateau” suggests, these features ther the lava volatile content nor the effusion
are typically 1–3 km higher than the surround- rates were sufficiently high to generate explo-
ing oceanic floor; geophysical data suggest that sions during their genesis. The extraordinary
Figure 12.2╇╇ A perspective view north from 7°N along Figure 12.3╇╇ Same bathymetric map as shown in Fig. 12.2,
the East Pacific Rise. Note how the two east–west trending with first-, second-, and third-order discontinuities labeled.
transform faults offset the ridge near the center of the image. See http://media.marine-geo.org; modified from Ryan et€al.
See http://media.marine-geo.org; modified from Ryan et€al. (2009).
(2009).
Table 12.1╇ ╇ Average major element analyses of individual lava flows from mid-ocean ridges.
Spreading rate
Spreading rate (cm/yr, full rate) Type locale Citation
Ultra-slow 0.6–1.3 Gakkel Ridge Sohn et€al. (2008)
Slow 1.0–4.0 Mid-Atlantic Ridge Perfit and Chadwick (1998)
Intermediate 4.0–8.0 Juan de Fuca Ridge Perfit and Chadwick (1998)
Fast 8.0–16.0 East Pacific Rise, 9°50’N Perfit and Chadwick (1998)
Superfast >16.0 East Pacific Rise, 30°S Korenaga and Hey (1996)
example, dacites have been found along the to correlate with spreading rate. Herein, the
submarine MOR system (Stakes et€ al., 2006; spreading rates given are full spreading rates: if
Wanless et€ al., 2010), but basalt€ – specifically, you were to stand on one plate, the full spread-
mid-ocean ridge basalt (MORB) is the most abun- ing rate is how fast the plate on the other side
dant lava erupting from submarine mid-ocean of the MOR would appear to be moving away
ridges. Mid-ocean ridge basalts are classified from you. Spreading rates vary from ultraslow
by their major- and trace-element abundances. (<1.5 cm/yr) to superfast (>16 cm/yr) (Table 12.2).
The range of Mg abundance in MOR basalts The Southern East Pacific Rise (SEPR) near
can vary considerably in time and space, as can 17°30’ S is the type locale for a superfast spread-
Ti and others (see Table 12.1). Closely spaced ing center (e.g., Sinton et€al., 2001); the Gakkel
(1–2 km) sampling along and across MORs and Ridge in the Arctic Ocean is the type locale for
subsequent analyses reveal that ridge discon- an ultraslow spreading center (Edwards et€ al.,
tinuities can coincide with significant changes 2001).
in underlying magma chemistry (Langmuir Eruptions along fast-spreading MORs appear
et€al., 1986; Sinton et€al., 1991, 2001; Perfit and to be small volume (~106–107 m3), frequent
Chadwick, 1998). (approximately every 5–10 years per kilometer
Eruption parameters, including erupted vol- of ridge length) and short duration (hours)
ume, eruption duration, and effusion rate (here, (Gregg et€al., 1996; Sinton et€al., 2001; Chadwick,
defined as volumetric rate, or m3 s−1), are believed 2003; Soule et€al., 2007). In contrast, eruptions at
slow spreading centers are thought to be larger ambient conditions. The high ambient pres-
volume and longer-lived (Perfit and Chadwick, sure at MORs (250 MPa) keeps most volatiles
1998; Sinton et€al., 2001). dissolved within the magma; a typical vesicle
content for MOR basalts is <2 vol.% (e.g., Perfit
and Chadwick, 1998), although volumes as high
as 35 vol.% have been reported (Hekinian et€al.,
12.3╇ ╇╇ Effusive submarine eruptions 2000). Commonly, however, volatile contents
of MORB magmas are thought to be much less
In 1977, R. Ballard and T. Van Andel led an than 0.5 wt.% (Byers et€al., 1986). For example, it
expedition to the Mid-Atlantic Ridge near 37°N is unlikely that true a’a lavas are produced on
(Ballard and van Andel, 1977). In addition to the deep sea floor because the required volatile
scholarly journal articles, Ballard and Moore exsolution (and accompanied increase in lava
(1977) published a “coffee-table” book on the viscosity) will not occur on the deep sea floor
expedition, complete with striking full-color as it does on land. Different descriptive terms
photographs of the deep ocean floor. By chance, are therefore used for submarine lava morph-
the region they explored was covered with pil- ologies, even though they may look similar to
low lavas (Ballard and van Andel, 1977; Ballard subaerial basalt morphologies. Submarine lava
and Moore, 1977); thereafter, most introductory flows are described as pillowed, lobate, lineated,
geology textbooks stated that the ocean floor ropy (folded), or jumbled (hackly) (Fig. 12.4).
was dominated by pillow lavas. Furthermore,
pillowed lavas were observed to form under- 12.3.1â•… Submarine lava flow inflation
water where subaerial Hawaiian lavas flowed Observations suggest that almost all submar-
into the ocean (Tepley and Moore, 1974; Tribble, ine lava flows are inflated (Hon et€ al., 1994) to
1991; see online supplement OS12B.1, “Pele some extent, including pillow lavas. If lava infla-
Meets the Sea”). In addition, subglacial lavas tion is defined as “molten lava injected beneath
also commonly display pillowed morphologies a solidified lava crust,” then the rapidity with
(e.g., Jones, 1970; Jakobsson and Gudmundsson, which submarine basalts form solidified crusts
2008). The presence of lava pillows, therefore, helps to explain the abundance of lava inflation
requires an abundance of water, but an abun- features on the seafloor (Applegate and Embley,
dance of water does not produce only lava pil- 1992; Gregg and Chadwick, 1996; Chadwick
lows (cf. Embley et€al., 1991): most of the Pacific et€al., 1999). Submarine inflation features similar
Ocean is likely floored by a type of sheet flow to those observed on land€– such as tumuli€– have
known as lobate (Fox et€ al., 1988; Perfit and been identified (Applegate and Embley, 1992).
Chadwick, 1998; Fig. 12.4). Submarine lava pillars (Fig. 12.5) are also indi-
On the deep sea floor (most MORs have a cative of lava inflation. They are hollow cylinders
water depth of ~2500 m; Perfit and Chadwick, of basalt, morphologically similar to subaerial
1998), steam formation and explosions driven lava trees or tree molds. They are commonly < 20 m
by magmatic gases are minor or rare (cf. Perfit tall and < 2 m in diameter, and form during the
et€al., 2003; Clague et€al., 2009). At a water depth emplacement, inflation, and drainback of lobate
of 2500 m, the ambient pressure is 250 MPa. sheet flows in submarine environments (Gregg
A normal MORB would theoretically require and Chadwick, 1996; Chadwick et€ al., 1999;
a water content of 4.5 wt.% H2O before a mag- Chadwick, 2003). Gregg and Chadwick (1996)
matic explosion could occur, and most MORs do state that pillars grow as heated seawater rises
not contain sufficient volatiles (Table 12.1) for through gaps between adjacent lava lobes of an
magma fragmentation (Head and Wilson, 2003). advancing submarine lava flow. Alternatively,
Therefore, most MOR eruptions are effusive. it has been proposed that pillars form after the
Submarine lava flow morphologies are lava has been emplaced, when buoyant water
similar to, but distinct from, subaerial basaltic rises through the overlying lava (Francheteau
lava morphologies because of the different et€al., 1979; Perfit et€al., 2003). Cooling at the flow
(Tc − Ta ) 3
4
was thin enough to flow beneath the 50-cm-tall Fconv = ρa c a γ (12.1)
legs of the instrument. Subsequently, the lava νa
flow inflated and then drained away, leaving a
lava flow only ~1 m thick. The entire inflation/ where Fconv is the convective heat flux from the
deflation event took only 153 minutes (Fox et€al., lava flow surface (W m−2); ρa is density (kg m−3) of
2001). Lava pillars about 2 m tall were subse- the ambient medium (here, seawater at 2500€m
quently found within the lava flow (Chadwick, depth and 4°C); ca is seawater specific heat
2003). (J kg−1 K−1); γ is a dimensionless constant equal
Similarly, lobate lava flows along the East to 0.1; g′ is reduced gravitational acceleration
Pacific Rise are commonly pock-marked with (m s−2); αa is the coefficient of thermal expan-
collapse pits, revealing hollow interiors locally sion of seawater (K−1); κa is seawater thermal
supported by lava pillars (Fornari et€ al., 1998; diffusivity (m2 s−1); νa is the seawater kinematic
Sinton et€al., 2001; Perfit et€al., 2003; Soule et€al., viscosity (m2 s−1); Ta and Tc are the temperatures
2006; Fundis et€al., 2010). Observations therefore of the seawater and lava crust (K). (Notation is
indicate that lobate flows are emplaced by lava summarized in Section 12.8.) Reduced gravity,
inflation processes. g′, takes into account the density of the sur-
rounding seawater, and is given by (Fink and
Griffiths, 1990):
g ′ = g (ρ − ρa ) / ρ
12.3.2╇ ╇╇Modeling effusive (12.2)
Fink, 1992). Videos taken of subaerially erupted is given by (Fink and Griffiths, 1990; Griffiths
lavas on Kilauea volcano that flowed into the and Fink, 1992):
ocean confirm these calculations (Tepley and 3
ν 4
− 14 (12.5a)
Moore, 1974; see OS12B.1). Additionally, MOR ta = l Q
g′
lavas are typically covered with a glassy rind
≤2 cm thick, confirming rapid cooling to for a point-source eruption, and by:
the glass transition temperature (Gregg and
2
Chadwick, 1996, and references therein). ν 3
− 13
ta = l q (12.5b)
The glassy rind on MOR lavas acts to insu- g′
late the interior of the lava flow from add-
itional heat loss. Heat from the lava interior for a line-source (fissure) eruption in which q (m2
can only be transferred to the lava flow surface s−1) is the volumetric effusion rate per meter of
via conduction through this glassy rind; and fissure length. Most mid-ocean ridge eruptions
conduction is a slow mechanism of heat trans- appear to be fissure eruptions (Chadwick and
fer (see Chapter€ 5). Once the glass rind forms, Embley, 1998; Fornari et€ al., 1998). Laboratory
the lava flows erupted from MORs tend to be simulations, in which polyethylene glycol (PEG)
volume-limited rather than cooling-limited was extruded at a constant effusion rate with a
(Guest et€ al., 1987; Gregg and Fornari, 1998). controlled cooling rate, allowed Ψ values to be
Thus, in comparison with subaerial basalt flows, quantitatively attached to PEG flow morpholo-
submarine lavas: (1) contain fewer vesicles; (2) gies. These morphologies were related to those
display a glassy rind ≥2 cm thick formed by sea- seen at mid-ocean ridges (Gregg and Fink, 1995)
water rapidly quenching their surface; (3) have to constrain eruption rates there (Fig. 12.6).
interiors that cool more slowly because they are Using this method, Gregg et€al. (1996) estimated
insulated by the glassy rind. eruption durations for MOR eruptions; the 1998
Although the pressure exerted at the great eruption at Axial Volcano on the Juan de Fuca
depths of MORs suppresses steam explosions, Ridge confirmed this methodology (Chadwick,
there is still evidence for vapor pockets exist- 2003).
ing in contact with submarine lavas (Perfit If eruption viscosity is known or can be esti-
et€ al., 2003; Chadwick, 2003; Soule et€ al., 2006; mated, and flow morphology assigned to one of
Schiffman et€al., 2010). These pockets are small the categories identified by Fink and Griffiths
(tens of cm across or smaller) and short-lived, (1990) or Gregg and Fink (1995), and cooling
but nonetheless contribute to the formation of rates calculated, then effusion rates can be con-
cm- to nm-scale features observed on and in sub- strained. Griffiths and Fink (1992), Gregg and
marine lava flow crusts. Greeley (1993) and Gregg and Fink (1996) dem-
Fink and Griffiths (1990) developed a dimen- onstrated the applicability of this technique for
sionless parameter, Ψ, that can be related to extraterrestrial lava flows, for example.
lava flow morphology if the eruption rate, Q, Klingelhofer et€al. (1999) examined the fluid
and lava kinematic viscosity, νl, are known. Ψ dynamic behavior of cooling basalt with a
is essentially the ratio of the timescale required temperature-dependent viscosity to constrain
for the surface of a lava flow to solidify (ts) to the formation of submarine lava pillows. They
the timescale for heat to be advected along the modeled pillows as cylindrical “tubes” with
flow (ta): known dimensions, and concluded that the
cooling rate must be balanced by the advection
ts
Ψ= . of hot lava, which in turn is controlled by a pres-
ta (12.4)
sure gradient. The critical pressure gradient can
The timescale for solidification is proportional be related to the radius (r) of the pillow tube
to the heat flux from the lava flow surface; raised to the power 3.85 (i.e., r3.85). Consistent
on the seafloor, this is dominated by convec- with the laboratory simulations of Fink and
tion (Eq. (12.1)). The timescale for advection Griffiths (1990), Klingelhofer et€ al. (1999) agree
limu o Pele or Pele’s seaweed) along the East Pacific Embley et€al. (2006) report on the discovery of
Rise, the Juan de Fuca and Gorda Ridges, at Loihi eruptive activity at NW Rota-1, a basaltic to bas-
seamount, along the Hawaiian arch, and in the altic andesite volcano located 60 km northwest
Fiji back-arc basin. Sohn et€al. (2008) discovered of Rota island. The volcano’s summit is ~500 m
limu o Pele deposits at 4000 m depth along the below sea level. This find marks the first direct
Gakkel ridge. Clague et€ al. (2009) found a cor- observations and sampling of a submarine arc
relation between the abundance of limu o Pele volcano, and provides opportunities for investi-
and “high-rate eruptions,” identified on the gating the behavior of deep submarine volcanic
basis of lava flow morphologies: higher-effusion explosions (Chadwick et€ al. 2008a; in review).
rate lava flow morphologies are associated with During explosive events, the volcano emitted
a greater abundance of bubble wall fragments sulfur-rich eruption plumes, bubbles of CO2,
than are lower-effusion rate morphologies. They and basaltic andesite scoria (Embley et€al., 2006).
assert that the morphology and composition of Deardorff et€al. (2011) report on a week’s worth
these limu o Pele require that they formed from of explosive and effusive activity at the submar-
violent expansion of magmatic gases (probably ine volcano NW Rota-1 in 2006. Activity ranged
CO2; Head and Wilson, 2003; Sohn et€al., 2008), from diffuse gas venting to energetic explosions
rather than from explosions caused by expand- (see online supplement OS12B).
ing seawater, and suggest that the fragmenta- Eruption plumes at NW Rota-1 had both
tion and dispersal of limu o Pele are consistent momentum-driven and buoyancy-driven com-
with a strombolian eruption style. Similar to ponents. Video analyses show that momentum-
models presented by Head and Wilson (2003), driven plumes rose with velocities from ~1–4
Clague et€al. (2009) propose that the submarine m€s−1, whereas the buoyancy-driven plumes rose
strombolian activity was generated by the add- at a steady ascent rate of ~0.35 m s−1. Based on
ition of a magmatic foam from the top of the Wilson and Self’s (1980) model, Deardorff et€al.
magma reservoir to the resident magma. (2011) estimated the efficiency of plume cooling
Schipper et€ al. (2010) collected lapilli and by seawater by considering the plume ascent
bombs from near the summit of Loihi sea- rate:
mount, at water depths of 1100–1291 m. They 8rg ∆ρ
u= (12.6)
examined vesicle textures, as well as the geo- 3ρa Cd
chemistry of the lavas, and the H2O and CO2
contents of the lapilli glasses, ash, and glass in which u is the velocity of plume rise (m
inclusions. Vesicles composed 34–45 vol.% of the s−1); r is the plume radius (measured from
samples examined€– well below the fragmenta- video); Cd is the drag coefficient (assumed to
tion volume for subaerial eruptions (Wilson, be 0.5 for spherical particles); and ∆ρ is the
1980). Vesicles were poorly connected, in spite density contrast between plume and seawater
of their abundance. These characteristics are (kg€m−3). Deardorff et€al. (2011) used measured
similar to basaltic �scoria generated from sub- velocities and plume radii, combined with
aerial hawaiian-style (lava fountain) eruptions. known thermophysical properties of seawater
The ~1 km water depth corresponds to an ambi- (Fofonoff and Millard, 1983), to calculate the
ent pressure of ~10 MPa. At this pressure, the density contrast ∆ρ, and then to relate ∆ρ to
presence of H2O can affect the solubility of a required temperature contrast (neglect-
CO2 in basalt (see Chapter 4), and this volatile ing the effect of volcanic particles on plume
coupling affects the resulting eruption style. density). Observations of incandescence at the
Schipper et€al. (2010) assert that hydromagmatic vent requires temperatures of at least 700 °C
fragmentation likely occurs below the vent level (Chadwick et€ al., 2008a), and the calculations
of most submarine volcanoes, and that a strong require temperatures on the order of 30° C
coupling of volatile-phase bubbles with ris- within the plume at ~1 m above the vent. This
ing magma may be a requirement for efficient implies that efficient, rapid cooling (>200° C
magma fragmentation. s−1) of the plume occurs.
Cochran, J. R. (2008). Seamount volcanism along East Pacific Rise 0°Â€– 24°N based on beam data.
the Gakkel Ridge, Arctic Ocean. Geophysics Journal In Seamounts, Islands and Atolls, ed. B. H. Keating,
International, 174, 1153–1173. P. Fryer, R. Batiza and G. W. Boehlert. American
Coffin, M. F. and O. Eldholm (1994). Large igneous Geophysical Union Geophysical Monograph 43,
provinces: Crustal structure, dimensions, and Washington, DC, pp. 13–24.
external consequences. Reviews of Geophysics, 32, Fornari, D. J., Haymon, R. M., Perfit, M. R., Gregg,
1–36. T. K. P. and Edwards, M. H. (1998). Axial summit
Deardorff, N. D., Cashman, K. V. and Chadwick, trough of the East Pacifi Rise 9°-10°N; Geological
W. W., Jr. (2011). Observations of eruptive characteristics and evolution of the axial zone
plume dynamics and pyroclastic deposits from on fast spreading mid-ocean ridge. Journal of
submarine eruptions at NW Rota-1, Mariana arc. Geophysical Research, 103(B5), 9827–9855.
Journal of Volcanology and Geothermal Research, 202, Fox, C. G. (1999). In situ ground deformation
47–59. measurements from the summit of Axial Volcano
Dziak, R. P. and Fox, C. G. (1999). The January 1998 during the 1998 volcanic episode. Geophysical
earthquake swarm at Axial Volcano, Juan de Research Letters, 26, 3437–3440.
Fuca Ridge: Hydroacoustic evidence of seafloor Fox, C. G., Murphy, K. M. and Embley, R. W. (1988).
volcanic activity. Geophysical Research Letters, 26, Automated display and statistical analysis of
3429–3432. interpreted deep-sea bottom photographs. Marine
Edwards, M. H., Kurras, G. J., Tolstoy, M. et€al. Geology, 78, 199–216.
(2001). Evidence of recent volcanic activity on the Fox, C. G., Chadwick, W. W., Jr. and Embley, R. W.
ultraslow-spreading Gakkel ridge. Nature, 409, (2001). Direct observation of a submarine volcanic
808–812. eruption from a sea-floor instrument caught in a
Embley, R. W., Chadwick, W. W., Jr., Perfit, M. R. lava flow. Nature, 412, 727–729.
and Baker, E. T. (1991). Geology of the northern Francheteau, J., Juteau, T. and Rangan, C. (1979).
Cleft segment, Juan de Fuca Ridge: Recent lava Basaltic pillars in collapsed lava-pools on the deep
flows, sea-floor spreading, and the formation of ocean floor. Nature, 281, 209–211.
megaplumes. Geology, 17, 771–775. Fundis, A. T., Soule, S. A., Fornari, D. J. and
Embley, R. W., Chadwick, W. W., Jr., Baker, E. T., et€al. Perfit, M. R. (2010). Paving the seafloor: Volcanic
(2006). Long-term eruptive activity at a submarine emplacement processes during the 2005–2006
arc volcano. Nature, 441, 494–497. eruptions at the fast spreading East Pacific Rise,
Embley, R. W., Baker, E. T., Butterfiled, D. A. et€al. 9°50’N. Geochemistry Geophysics Geosystems, 11,
(2007). Exploring the submarine Ring of Fire. doi:10.1029/2010GC003058.
Oceanography, 20, 68–79. Gregg, T. K. P. and Chadwick, W. W., Jr. (1996).
Fink, J. H. and Griffiths, R. W. (1990). Radial Submarine lava-flow inflation: A model for the
spreading of viscous-gravity currents with formation of lava pillars. Geology, 24, 981–984.
solidifying crust. Journal of Fluid Dynamics, 221, Gregg, T. K. P. and Fink, J. H. (1995). Quantification
485–509. of submarine lava flow morphology through
Fink, J. H. and Griffiths, R. W. (1992). A laboratory analog experiments. Geology, 23, 73–76.
analog study of the surface morphology of lava Gregg, T. K. P. and Fink, J. H. (1996). Quantification
flows extruded from point and line sources. of extraterrestrial lava flow effusion rates
Journal of Volcanology and Geothermal Research, 54, through laboratory simulations. Journal of
19–32. Geophysical Research, 101(E7), 16891–16900.
Fofonoff, P. and Millard, R. C., Jr. (1983). Algorithms Gregg, T. K. P. and Fink, J. H. (2000). A laboratory
for computation of fundamental properties of investigation into the effects of slope on lava flow
seawater. UNESCO Technical Papers in Marine Science, morphology. Journal of Volcanology and Geothermal
44, 1–53, http://www.es.flinders.edu.au/~mattom/ Research, 96, 281–292.
Utilities/density.html. Gregg, T. K. P. and Fornari, D.J. (1998). Long
Fornari, D. J. (2003). A deep-sea towed digital camera submarine lava flows: Observations and results
and multi-rock coring system. Eos, Transactions of from numerical modeling. Journal of Geophysical
the American Geophysical Union, 84(8), 69–76. Research, 103(B11), 27517–27531.
Fornari, D. J., Batiza, R. and Luckmann, M. A. (1987). Gregg, T. K. P. and Greeley, R. (1993). Formation of
Seamount abundances and distribution near the venusian canali: Considerations of lava types
and their thermal behaviors. Journal of Geophysical Jones, J. G. (1970). Intraglacial volcanoes of the
Research, 98(E6), 10 873–10 882. Laugarvatn region, southwest Iceland II. Journal of
Gregg, T. K. P., Fornari, D. J., Perfit, M. R., Haymon, Geology, 78, 127–140.
R. M. and Fink, J. H. (1996). Rapid emplacement Klingelhofer, F., Hort, M., Kumple, J.-J. and
of a mid-ocean ridge lava flow on the East Pacific Schmincke, H.-U. (1999). Constraints on the
Rise at 9°46’-51’N. Earth and Planetary Science formation of submarine lava flows from
Letters, 144, E1–E7. numerical model calculations. Journal of
Griffiths, R. W. and Fink, J. H. (1992). Solidification Volcanology and Geothermal Research, 92, 215–229.
and morphology of submarine lavas: A Korenaga, J. and Hey, R. (1996). Recent dueling
dependence on extrusion rate. Journal of propagation history at the fastest spreading
Geophysical Research, 97(B13), 19 729–19 737. center, the East Pacific Rise, 26°–32°S. Journal of
Guest, J. E., Kilburn, C. R. J., Pinkerton, H. and Geophysical Research, 101(B8), 18 023–18 041.
Duncan, A. M. (1987). The evolution of lava Langmuir, C. H., Bender, J. F. and Batiza, R. (1986).
flow-fields: Observations of the 1981 and 1983 Petrological and tectonic segmentation of the
eruptions of Mount Etna, Sicily. Bulletin of East Pacific Rise, 5°30’–14°30’N. Nature, 322,
Volcanology, 49, 527–540. 422–429.
Haymon, R. M., Fornari, D. J., von Damm, K. L., Macdonald, K. C. (1998). Faulting, volcanism,
et€al. (1993). Volcanic eruption of the mid-ocean hydrothermal activity on fast spreading centers,
ridge along the East Pacific Rise crest at 9°45’- in Faulting and Magmatism at Mid-ocean Ridges.
52’N: Direct submersible observation of seafloor American Geophysical Union Geophysical
phenomena associated with an eruption event in Monograph 106, Washington, D.C., pp. 27–58.
April 1991. Earth and Planetary Science Letters, 119, Neal, C. R., Mahoney, J. J., Kroenke, L. W., Duncan,
85–101. R. A. and Petterson, M. G. (1997). The Ontong
Head, J. W. and Wilson, L. (2003). Deep submarine Java Plateau. In Large Igneous Provinces: Continental,
pyroclastic eruptions: Theory and predicted Oceanic, and Planetary Flood Volcanism, ed. J. J.
landforms and deposits. Journal of Volcanology and Mahoney and M. F. Coffin. Washington, DC:
Geothermal Research, 121, 155–193. American Geophysical Union Geophysical
Hekinian, R., Pineau, F., Shilobreeva, S. et€al. Monograph 100, pp. 183–216.
(2000). Deep sea explosive activity on the Ocean Drilling Program (1991). Ontong Java Plateau,
Mid-Atlantic Ridge near 34°50’N: Magma Scientific Results of the Ocean Drilling Program, 130.
composition, vesicularity, and volatile content. Ocean Drilling Program (2001). Basement drilling
Journal of Volcanology and Geothermal Research, 98, of the Ontong Java Plateau, Scientific Results of the
49–77. Ocean Drilling Program, 192.
Hon, K., Kauahikaua, J., Denlinger, R. and Mackay, K. Perfit, M. R. and Chadwick, W. W., Jr. (1998).
(1994). Emplacement and inflation of pahoehoe Magmatism at mid-ocean ridges: Constraints from
sheet flows: Observations and measurements volcanological and geochemical investigations.
of active lava flows on Kilauea Volcano, Hawaii. In Faulting and Magmatism at Mid-ocean Ridges.
Geological Society of America Bulletin, 106, 351–370. Washington, DC: American Geophysical Union
Humphris, S. E., Bryan, W. B., Thompson, G. and Geophysical Monograph 106, pp. 59–115.
Autio, L. K. (1990). Morphology, geochemistry and Perfit, M. R., Fornari, D. J., Smith, M. C. et€al. (1995).
evolution of Serocki volcano. In Proceedings of the Small-scale spatial and temporal variations
Ocean Drilling Program, Scientific Results, 106/109, in mid-ocean ridge crest magmatic processes.
ed. R. S. Detrick, J. Honnorez, W. B. Juteau et€al. Geology, 22, 375–379.
College Station, Texas: Ocean Drilling Program, Perfit, M. R., Cann, J. R., Fornari, D. J. et€al. (2003).
pp. 67–84. Interaction of sea water and lava during
Ito, G., Lin, J. and Graham, D. (2003). Observational submarine eruptions at mid-ocean ridges. Nature,
and theoretical studies of the dynamics of mantle 426, 62–65.
plume-mid-ocean ridge interaction. Reviews of Rubin, K. H., Smith, M. C., Bergmanis, E. C. et€al.
Geophysics, 41, 1017, doi:10.1029/2002RG000117. (2001). Magmatic history and volcanological
Jakobsson, S. P. and Gudmundsson, M. T. (2008). insights from individual lava flows erupted on the
Subglacial and intraglacial volcanic formations in sea floor. Earth and Planetary Science Letters, 188,
Iceland. Jokull, 58, 179–196. 349–367.
Rubin, K. H., Embley, R. W., Clague, D. A. et€al. glycol 600, and analog for lava crust. Journal
(2009). Lavas from active boninite and very of Volcanology and Geothermal Research, 129,
recent basalt eruptions at two submarine NE 139–153.
Lau Basin sites. Eos, Transactions of the American Soule, S. A., Fornari, D. J., Perfit, M. R. et€al.
Geophysical Union, 89, Fall Meeting Supplement, (2005). Channelized lava flows at the East
abstract V43I-05. Pacific Rise Crest 9°–10°N: The importance
Ryan, W. B. F., Carbotte, S. M., Coplan, J. et€al. of off-axis lava transport in developing
(2009). Global multi-resolution topography the architecture of young oceanic crust.
(GMRT) synthesis data set. Geochemistry Geochemistry Geophysics and Geosystems, 16,
Geophysics and Geosystems, 10, Q03014, doi:10.1029/2005GC000912.
doi:10.1029/2008GC002332. Soule, S. A., Fornari, D. J., Perfit, M. R. et€al. (2006).
Schiffman, P., Zierenberg, R., Chadwick, W. W., Jr., Incorporation of seawater into mid-ocean ridge
Clague, D. A. and Lowenstern, J. (2010). lava flows during emplacement. Earth and
Contamination of basaltic lava by seawater: Planetary Science Letters, 252, 289–307.
Evidence found in a lava pillar from Soule, S. A., Fornari, D. J., Perfit, M. R. and Rubin, K.
Axial Seamount, Juan de Fuca Ridge. (2007). New insights into mid-ocean ridge volcanic
Geochemistry Geophysics and Geosytems, 11, processes from the 2005–2006 eruption of the
doi:10.1029/2009GC003009. East Pacific Rise, 9°46’N–9°56’N. Geology, 35,
Schipper, C. I., White, J. D. L., Houghton, B. F., 1079–1082.
Shimizu, N. and Stewart, R. B. (2010). “Poseidic” Soule, S. A., Escartin, J. and Fornari, D. J.
explosive eruptions at Loihi Seamount, Hawaii. (2009). A record of eruption and intrusion
Geology, 38, 291–294. at a fast spreading ridge axis: Axial summit
Self, S., Kesztheyli, L. P. and Thordarson, Th. (1998). trough of the East Pacific Rise at 9–10°N.
The importance of pahoehoe. Annual Review of Geochemistry Geophysics and Geosystems, 10,
Earth and Planetary Sciences, 26, 81–110. doi:10.1029/2008GC002354.
Sinton, J. M. and Detrick, R. S. (1992). Mid-ocean Stakes, D. S., Perfit, M. R., Tivey, M. A. et€al. (2006).
ridge magma chambers. Journal of Geophysical The Cleft revealed: Geologic, magnetic, and
Research, 97(B1), 197–216. morphologic evidence for construction of upper
Sinton, J. M., Smaglik, S. M., Mahoney, J. J. and oceanic crust along the southern Juan de Fuca
Macdonald, K. C. (1991). Magmatic processes Ridge. Geochemistry Geophysics and Geosystems, 7,
at superfast spreading mid-ocean ridges: glass doi:1029/2005GC001038.
compositional variations along the East Pacific Stern, R. J., Tamura, Y., Leybourne, M. I. et€al.
Rise, 13°-23°S. Journal Geophysical Research, 96, (2010). Felsic magmatism in intra-oceanic arcs:
6133–6155. The Diamante Cross-chain in the Southern
Sinton, J. M., Bergmanis, E., Rubin, K. et€al. (2001). Mariana Arc, Abstract #V14A-02, presented at
Volcanic eruptions on mid-ocean ridges: New 2010 American Geophysical Union Fall Meeting, San
evidence from the superfast spreading East Pacific Francisco, CA, 13–17 December.
Rise, 17°S–19°S. Journal of Geophysical Research, Tamura, Y., Ishizuka, O., Stern, R. J. et€al. (2010). Two
107(B6), doi:10.1029/2000JB000090. primary magma types from Northwest Rota-1
Smith, D. K. and Cann, J.R. (1992). The role of Volcano, Mariana Arc, Abstract #T13B-2197,
seamount volcanism in crustal construction presented at 2010 American Geophysical Union Fall
at the mid-Atlantic Ridge. Journal of Geophysical Meeting, San Francisco, CA, 13–17 December.
Research, 97(B2), 1645–1658. Tepley, L. and Moore, J. G. (1974). Fire Under the Sea:
Smithsonian Institution (2011). Global Volcanism The Origin of Pillow Lava. Mountain View, CA:
Program, www.volcano.si.edu/info/about/about_ Moonlight Productions.
votw.cfm. Tolstoy, M., Cowen, J. P., Baker, E. T. et€al. (2006).
Sohn, R. A., Willis, C., Humphris, S. et€al. (2008). A seafloor spreading event captured by
Explosive volcanism on the ultraslow-spreading seismometers. Science, 314, 1920–1922.
Gakkel Ridge, Arctic Ocean. Nature, 453, Torresan, M. E., Clague, D. A. and Jacobs, C. L. (1991).
doi:10.1038/nature07075. Cruise report, Hawaiian GLORIA Cruise F12–
Soule, S. A. and Cashman, K. V. (2004). The 89-HW. United States Geological Survey, Open-File
mechanical properties of solidified polyethylene Report #67.
Volcano–ice interactions
Lionel Wilson, John L. Smellie, and James W. Head
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
escapes in a sudden, rapid, and environmen- Case (a), injection of magma into an ice layer
tally devastating discharge known as a glacial as a dike, is predicted to be a possibility on the-
outburst flood or jökulhlaup. Several of these fea- oretical grounds (Wilson and Head, 2002). The
tures were displayed during the exceptionally inherent instability of the resulting geometry
well-documented eruption at Gjálp, Iceland, in means that a fragmental deposit surrounding
1996, which took place under the 700-m-thick the vent (fissure) at the base of the ice sheet is
Vatnajökull ice cap (Gudmundsson et€al., 1997, likely to be produced at the end of the inter-
2002, 2004; Jarosch et€ al., 2008). Although action. The subsequent evolution of the subgla-
the spectacular April–May 2010 eruption at cial volcanic mass might involve construction
Eyjafjallajökull (also Iceland) took place under of a pillow mound, then an explosively gener-
much thinner ice (just 200–300 m), it also dem- ated subaqueous tuff cone and ultimately a sub-
onstrated many similar aspects of volcano–ice aerial edifice with or without lava-fed deltas, as
interactions, in particular a high explosively described for many glaciovolcanic edifices (e.g.,
generated tephra-bearing eruption column and Jones, 1969, 1970; Smellie, 2000). Some of these
meltwater floods that washed away the local eruptions might have been associated with
highway. Unlike the Gjálp eruption, the effects jökulhlaups.
of which were limited to Iceland and included In case (b), melting of ice overlying a sill
the creation of new land as well as serious and intrusion at the ice base can produce a large vol-
costly damage to the local transport infrastruc- ume of water (Höskuldsson and Sparks, 1997;
ture, the Eyjafjallajökull eruption distributed Wilson and Head, 2002, 2007; Tuffen, 2007). The
ash over much of northern Europe. It is esti- water may accumulate relatively slowly, but in
mated that about $1.7 billion was lost to the the case of a large body of water accumulated
EU economy as a result principally of flights under a glacier it may eventually escape very
that had to be cancelled (see Eyjafjallajökull rapidly in the form of a jökulhlaup. If the water
links at end of chapter). In this chapter the escapes at the glacier edge and is replaced by
human impact of volcano–ice eruptions is not air, and the overlying ice does not deform too
described. Rather, the focus is on providing a quickly, the sill may evolve into what is effect-
theoretical framework of how mafic magmas ively a subaerial lava flow in an ice cave, result-
might interact with ice under different situ- ing in a range of possible features (Wilson and
ations, and examples are provided (where avail- Head, 2002, 2007). Key issues then are the rela-
able) that show what the resulting rock record tive values of the rate of advance of the magma
might look like. sheet beneath the ice, the rate of advance of
thermal waves penetrating both the unmelted
13.1.1â•…Three generic scenarios for ice and the magma itself, and the rate of drain-
volcano–ice interactions age of the meltwater produced. The evolution of
Ice, in the form of glaciers, snow-fall or frozen such an eruption is less well known but at least
lakes, commonly occurs as a sheet overlying a some might be associated with construction of
silicate rock surface. Magma penetrating the pillow mounds or ridges (Smellie, 2008).
shallow crust travels in brittle fractures, i.e., In case (c) the presence of the ice plays no
dikes and sills. Possible initial geometries of the part in the magmatic eruption. If the eruption
interaction between ice and magma are (a) injec- involves deposition of a pyroclastic fall deposit
tion of a magmatic dike into a layer of ice; (b) onto ice, the consequences depend on the rate
formation of what is initially a sill-like magma of thickening of the deposit and its final total
intrusion beneath ice at the ice–rock interface, thickness (e.g., Wilson and Head, 2007, 2009).
though this may ultimately develop into what Except very close to the vent, fall deposits are
might be thought of as a subglacial lava flow; likely to consist of clasts with temperatures
and (c) emplacement on top of ice of a lava flow close to the ambient atmospheric tempera-
or pyroclastic material originating elsewhere ture, and so immediate heating of underlying
(Wilson and Head, 2002, 2007, 2009). ice will be minimal. Longer term, the fact that
the pyroclasts will have a lower albedo than the the surface. Magma reservoirs are likely to form
ice, and hence will reach higher daytime tem- from the accumulation of multiple stalled intru-
peratures during the diurnal solar heating cycle, sions at intermediate levels of neutral magma
may be important. If the volcanic material is hot buoyancy (Ryan, 1987). Dikes are driven upward
on emplacement, e.g., a lava flow or a pyroclas- from these reservoirs by excess pressures, due to
tic density current deposit, the effects on the the positive buoyancy of the melts feeding them
underlying ice will depend on both the rate of from below, that are typically at least several
advance and thickness of the flowing material, MPa (Parfitt, 1991). Thus for reservoir depths
and the rate at which water can migrate beneath of ~3 km the pressure gradient in excess of the
the advancing deposit to escape at its edges. weight of the magma driving magma motion is
We now examine the above processes the- ~1000 Pa m−1. Where mafic magma reaches the
oretically, and suggest examples of the depos- surface directly from lower crustal or mantle
its that might form from them, to facilitate depths, net density differences between magma
the kinds of field observations needed to con- and host rocks are generally ~100 kg m−3 (Parfitt
strain the models and to identify issues so far et€al., 1993). This value, combined with the accel-
overlooked. eration due to gravity of ~10 m s−2, also leads to
pressure gradients of order 1000 Pa m−1. Widths
of mafic dikes propagating from shallow reser-
voirs are ~1–3 m (Parfitt, 1991), and pressure
13.2╇ ╇ Englacial dike emplacement gradients of this order drive mafic magma rise
speeds of ~1 m s−1 (Wilson and Head, 1981). The
13.2.1â•… Theoretical issues strain rates near the dike tips implied by these
Magma propagation at shallow depths in planet- speeds are ~1 s−1, about seven orders of magni-
ary bodies takes place in brittle fractures (dikes tude larger than the strain rates at which the
and sills), held open by a combination of magma surrounding ice can flow plastically given the
pressure and the stress field in the host rocks rheological models (a pseudo-plastic power-law
(Pollard, 1987; Rubin and Pollard, 1987; see fluid with a yield strength) proposed by Glen
Chapter 3). Some reach the surface to feed erup- (1952), Nye (1953) and Paterson (1994). Thus a
tions, but others stall as intrusions, and these mafic dike can easily overshoot an ice–rock inter-
may cause surface manifestations such as bed- face because the ice appears to the propagating
rock graben or ice fractures (crevasses) if their crack as a brittle, low-density rock with elas-
tops are sufficiently close to the surface (Mastin tic properties similar to those of the substrate
and Pollard, 1988; Rubin, 1992; Gudmundsson (Wilson and Head, 2002). The time required for
et€al., 2004). If the least principal stress changes a dike containing magma propagating at 1€m s−1
from horizontal to vertical, a dike will cease to to penetrate tens to hundreds of meters into ice
propagate upward but may still propagate lat- is tens to hundreds of seconds, and with both
erally to form a sill. A dike approaching the ice and rock having thermal diffusivities of
surface in a location capped by a substantial ~10−6 m2 s−1, the distance that heat can be con-
thickness of ice is likely to encounter such a ducted, chilling the magma and warming and
stress change which, coupled with a discontinu- then melting the ice, is ~20 mm. Thus the initial
ity in material properties, may encourage sill penetration of mafic dikes into ice may be a sta-
injection at the ice–rock interface. Such sills are ble process, though subsequent, more extensive
expected to evolve rapidly into subglacial lava ice melting may lead to collapse of the dike.
flows (Wilson and Head, 2002; Smellie, 2008; Consider a mafic dike that has propagated
see Section 13.3). from a magma reservoir whose top is at a depth
The density structure of the crust in volcanic z below an ice–rock interface (a list of all nota-
provinces is commonly such that mafic magma tion is given in Section 13.6). We adopt z = 3 km
generated in the mantle is positively buoyant as typical of many mafic reservoirs (Ryan, 1987).
in its source zone but negatively buoyant near The ice thickness is y (Fig. 13.1) and the top of
mafic pillow volcanoes, tindars, and tuyas (e.g., cease while the system still has this configur-
Jones, 1969, 1970; Skilling, 1994, 2002; Smellie ation, but heat will continue to be transferred
and Hole, 1997; Werner and Schmincke, 1999; from magma to ice, and the H2O volume decrease
Smellie, 2000, 2006; Schopka et€al., 2006). due to ice melting will at least partially relax
any excess pressure in the magma. If the glacier
is warm-based, water at the interface between
the glacier and the underlying rock will prevent
13.3╇ ╇Magma intrusion at the base the support of tensile stresses. The requirement
of an ice layer that the magma pressure at the dike-sill connec-
tion exceeds the weight of the glacier means
13.3.1â•… Theoretical issues that there is the potential for the glacial ice to
If a dike reaches an ice–rock interface and feeds be lifted by the magma. However, this can only
a sill intruding along the interface, the min- occur locally: ice has a finite shear modulus, and
imum requirement is that the pressure in the a thick and laterally extensive slab of ice (the
dike magma at the sill inlet must exceed the glacier) supported over an initially small region
pressure due to the weight of the overlying gla- beneath it (around the dike-sill connection) will
cial ice, Pg (Smellie, 2008). Thus, using Figure bend, remaining in contact with the underlying
13.1, we require rock beyond some finite distance from the sup-
ported region. As a result, in the early stages of
Pa + g (ρiy + ρrz) + ΔP€– (gρbz) > Pa + gρiâ•›yâ•… (13.4)
the intrusion the resulting stress in the ice will
which simplifies to interact with the magma pressure in the same
way as in the cold-based case, and the pressure
ΔP > gz (ρb€– ρr). (13.5)
in the intruding magma will be equal to or
Thus the required value of the excess magma greater than the weight of the overlying ice.
pressure ΔP increases with the depth of the res- However, as the sill margin approaches
ervoir z and the difference between the densities the edge of the glacier, a stage will be reached
of the dike magma, ρb, and the crustal rocks, ρr. when the glacial margin is lifted, providing a
The magma in the sill inlet will contain some local pressure connection to the atmosphere
exsolved CO2, but, except for a short time imme- and an escape route for water, water vapor, and
diately after the intrusion starts, the pressure at exsolved volcanic gases. Even in cold-based gla-
the inlet will not be as low as in the magma just ciers this condition must eventually be reached
behind the propagating sill tip where signifi- as the elastically propagating sill tip approaches
cant H2O has exsolved, and so a suitable magma the ice margin. Water loss is likely to happen
density is ~2600 kg m−3 (Wilson and Head, 2002). faster than the overlying ice can deform plastic-
With the value of ρr = 2300 kg m−3 adopted here, ally (Tuffen, 2007), especially if the water release
Eq. (13.5) then predicts that ΔP must be greater is on the scale of a jökulhlaup (e.g., Björnsson,
than ~9 MPa. The required value will be less 1992), so unless wholesale disintegration of the
than this in situations with more volatile-rich glacier occurs, air will progressively replace
or lower-density magma, denser crustal rocks, the water overlying the magma. The pressure
or shallower reservoirs. acting on the magma will be much closer to
A subglacial sill may evolve in a number of atmospheric pressure than to the weight of the
possible ways. If the glacier is cold-based, i.e., overlying glacier, and a wave of magma frag-
the ice temperature at the base of the ice layer mentation may propagate from the end of the
is less than the ice freezing point, the ice–rock sill near the ice margin both back toward the
interface will be strong enough to support ten- dike-sill inlet and downward into the sill as exist-
sile stresses and the intruding sill will propagate ing gas bubbles expand and additional volatiles
in the same way that a sill grows at a rock–rock exsolve from the magma. A possible example
interface, with both the ice and rock initially analogous to this is documented in Iceland by
behaving as elastic solids. The eruption may Höskuldsson et€al. (2006), in which early-formed
vesicular pillows underwent secondary vesicu- which it is emplaced, its volume effusion rate,
lation of the still-molten pillow cores, attrib- and its evolving rheological properties. The
uted to release of a jökulhlaup and associated absence of a free surface at the top of a sub-
sudden decompression of the pillows, although glacial magma body, together with the stress
fragmentation of the lava did not occur. In applied by the elastic properties of the ice above
extreme cases enough volatiles may be released, and rock beneath, will lead to it spreading into
especially from the least-cooled magma at the a sheet-like structure likely to be initially much
point where the feeder dike reaches the ice– wider, thinner, and more slowly advancing than
rock interface, that a lava fountain forms over a subaerial lava flow erupted at the same vol-
what becomes a subglacial fissure eruption. A ume flux. Other differences between subgla-
sufficiently high fountain will radiate heat to cial and subaerial behavior are similarly driven
the overlying ice (indeed, pyroclasts may even by the stress field. Thus, a subglacial intrusion
come into physical contact with it), thus greatly will get thicker as its edges advance, this pro-
enhancing ice melting in this region. We now cess offsetting the inward migration of the
elaborate on each of these three stages. effects of cooling at its upper and lower faces
and allowing it to advance to a much greater
13.3.2â•… Elastic intrusion distance from the vent than a subaerial flow
We argued above that the pressure in a sill before cooling limits its growth. The predicted
intruded beneath an ice layer must be at least thickening process has some similarities to the
equal to the pressure due to the weight of the inflation of subaerial lavas (Hon et€al., 1994) but,
overlying ice, and that the pressure in a propagat- whereas subaerial flow units inflate and thicken
ing dike or sill tip would probably not decrease after their margins come to rest as a result of
below the pressure at which magma fragmen- cooling, subglacial intrusions thicken continu-
tation occurs (a function of the total magma ously as their edges advance. These results will
volatile content and of order 3 MPa for a typical be broadly independent of magma composition,
mafic magma). Given that volatile exsolution in with caveats similar to those listed in Section
such a magma would be significant at pressures 13.2.1.
less that ~10 MPa (Chapter 4, Fig. 4.2), it is clear Heat transfer rates from subglacial magma to
that a large range of combinations of magma melted ice have been predicted by Höskuldsson
reservoir pressure and magma water and CO2 and Sparks (1997), Wilson and Head (2002, 2007),
contents are consistent with the injection of and Tuffen (2007), and measured in a few well-
vesicular magmatic liquids at ice–rock inter- studied cases (e.g., Jarosch et€al., 2008). The theor-
faces. Various aspects of the resulting growth etical calculations differ somewhat in the values
and evolution of the system have been explored used for material properties such as the specific
by Höskuldsson and Sparks (1997), Wilson and heat of the magma, which is typically averaged
Head (2002, 2007) and Tuffen (2007). over the temperature ranges involved, and the
Wilson and Head (2002) showed quantita- magma density, a function of composition and
tively that major morphological differences volatile content. Wilson and Head (2007) found
are expected between subaerial lava flows and that the volume of water that can be produced
elastically emplaced subglacial intrusions. The by ice melting is ~6.5 times as much as the
most obvious of these is the relative reduction intruded sill magma volume, which contrasts
of overall volatile exsolution due to the high with a volume ratio of ~10 if all of the magmatic
ambient pressure, in a manner analogous to heat is transferred to the water. Thus a 1-m-thick
deep submarine eruptions (Head and Wilson, sill would generate a water layer ~6.5 m deep by
2003; see Chapter 12). Other differences are melting a ~[(1000/900) × 6.5 =] ~7.2-m-thick layer
more subtle. A subaerial flow has an upper free of ice, the volume change thus accommodating
surface and acquires a cross-sectional shape (7.2€– 6.5 =) ~0.7 m of the 1 m sill thickness. As
(channel-and-levée, sheet-like, or compound suggested by Dixon et€ al. (2002) and by Figure
pahoehoe flow field) dictated by the slope on 4.2 in Chapter 4, as long as they do not make
pressure contact with the atmosphere, subgla- decrease to atmospheric pressure, and then
cial intrusions are expected to exsolve a sig- magma fragmentation must occur unless the
nificant fraction of their pre-eruption CO2 and total magma volatile content is extremely small
at most a minor fraction of their pre-eruption (Wilson and Head, 1981).
H2O. This suggests that examining residual vola- Magma fragmentation will begin at the mar-
tile contents of subglacially intruded rocks is a gin of the sill closest to the connection to the
potential diagnostic tool, in addition to purely atmosphere. As the pressure in the space above
morphological examination, for distinguishing the chilled magma crust decreases, the crust
such intrusions from subaerial lavas after all of may temporarily prevent any response from the
the overlying ice has been removed by climate underlying magma. However, the crust is likely
changes. Also, volatile concentrations found in to be pervaded by cooling cracks and have lit-
exposed subglacially erupted glasses can be used tle strength. When the crust fails, an expansion
(with many caveats) to reconstruct quenching wave will propagate down into the sill at a large
processes and minimum ice thicknesses (Dixon fraction of the local speed of sound which, in
et€al., 2002; Tuffen et€al., 2010). a vesicular liquid, will be ~100 m s−1 (Kieffer,
1977; Wilson and Head, 1981). Thus for a sill
13.3.3â•…Subglacial eruption after a few meters thick, this timescale will be less
atmospheric connection than a tenth of a second. The expansion wave
When the margin of a growing subglacial sill will fragment the magma (Scheu et€ al., 2006),
approaches sufficiently close to the edge of an and expansion of the released gas will accel-
ice sheet that a connection to the atmosphere is erate disrupted magma clasts to impact the
made, the first consequence will be the escape of overlying ice. Formulae given by Wilson (1980)
the pressurized water vapor in the sill tip. This for transient explosions show that, even if the
will be followed by progressive water leakage magma contains no water, a pressure reduction
and replacement by air, almost certainly leading from several MPa to 0.1 MPa in a magma still
to a wave of decompression advancing into the containing, say ~300 ppm of total dissolved vol-
sill from its distal margin towards the inlet from atiles will generate speeds in the hot vesiculated
the feeder dike. This can only be avoided com- pyroclasts of up to ~25 m s−1, thus projecting
pletely if the ice overlying the sill can deform them upward to heights of ~30 m. If the vertical
fast enough to replace the water and stay in distance to the base of the overlying ice is less
close proximity to the top of the sill magma. than this, the impact of these hot magma clasts
For conditions similar to those proposed here, will locally enhance ice melting; sufficiently
Höskuldsson and Sparks (1997) calculated an ice violent interaction of magma and ice may be
deformation rate of ~1 mm s−1, so if the rate of enough to trigger a sustained, violent molten
thinning of the water layer exceeds this value, fuel-coolant type of interaction (MFCI; Wohletz
some pressure decrease will inevitably occur. and McQueen, 1984; Zimanowski et€al., 1991; see
Initially, the magma will remain a vesicu- Chapter 11). The explosion products would be
lar foam: existing carbon dioxide bubbles will propelled toward the exit to the atmosphere as
expand and new bubbles of both CO2 and H2O the wave of magma vesiculation and fragmen-
will form at a rate that may allow the magma to tation propagated back toward the feeder dike.
stay in physical contact with the overlying ice. The propagation speed of the interaction would
If so, this will lead to continuing water produc- be a balance between the speed of the wave into
tion, and the system will tend toward an equilib- the as-yet unaffected sill magma (again some
rium where the pressure in the water is greatest large fraction of the ~100 m s−1 local speed of
near the dike and least at the edge of the ice sound) and the speed at which water and frag-
sheet, with the resulting pressure gradient driv- mented magma could be discharged from under
ing the water toward the exit. If water drainage the ice. Such an explosive fragmentation pro-
becomes efficient enough, it is possible that the cess is an excellent candidate for the origin of
pressure at the sill–ice contact will eventually sudden jökulhlaup production. Some fraction
of the fragmented magma would be washed out point of fragmentation. One or more discrete
with the escaping water and the rest would be lava flow units may eventually emerge from
left behind to form a vitroclastic (phreatomag- beneath the ice. After ice removal by climate
matic) deposit. change, it should be possible to identify this
When the wave of magma disruption reaches kind of flow by its changing shape as a function
the feeder dike, the system will behave like a of distance from its vent (taking due account of
subaerial eruption, with the formation of a chain pre-existing topography), and to distinguish it
of lava fountains. Depending on the efficiency from a flow generated by lava fountain activity
with which water is drained from the vicinity in an ice cavern after sill disruption.
of the vent, the lava fountains may be entirely
magmatic or may be phreatomagmatic. Initially 13.3.4â•…Geological examples of subglacial
the fountains will “drill” into the overlying ice. eruptions
A cavity will grow upward until the fountains Eruptions under thin ice, in cases where the
reach the subaerial height corresponding to the pressure is not enough to suppress magma
total magma volatile content (Head and Wilson, fragmentation, will be explosive and result in
1987). Thereafter heat will only be transferred the abrupt removal of overlying ice and con-
to the ice by radiation from pyroclasts in the struction of pyroclastic cones, as scoria or tuff
fountains, and the larger clasts falling from cones or tuff rings. Subglacial sills will not be
the fountains will begin to form rootless lava formed. The generation of extreme conditions
flows (Head and Wilson, 1989) spreading away such as formation of a pyroclastic density cur-
from the vent over the top of the disrupted sill rent (Section 13.3.3) and its emplacement by
material. A new balance will be approached intrusion beneath an ice cover is also unlikely
between ice subsidence and ice melting, and in these cases. Once magma fragmentation is
if the eruption continues for long enough the achieved by high levels of vesiculation, in the
explosive activity at the vent may eventually presence of abundant meltwater, a violent fuel–
emerge through the ice. Alternatively, the ice coolant interaction with meltwater will result
layer above the fragmented sill residue might in an abrupt transition to an explosive phreato-
fracture in a brittle fashion and collapse rather magmatic eruption (Chapter 11). This will form
than slowly deforming plastically. The pressure stratified tuff cones or tuff rings composed of
at the dike vent would still be close to atmos- interbedded fall and pyroclastic density cur-
pheric provided there was a reasonably high rent tephra. However, no examples have been
porosity and permeability in the collapsed ice observed, possibly because the edifice, formed
block pile, but interactions between magma of unconsolidated tephra, is easily eroded by
and ice would be more vigorous because of their ice. Additionally, some of the edifice will be
greater proximity. formed on the ice itself and will be destroyed
A hybrid situation could occur where the as the ice flows down slope and melts. The pyro-
magma volatile content is very low. If water clastic cones thus have a very low preservation
drainage is inefficient, the pressure reduction potential. However, distinctive glaciovolcanic
rate in the water above the sill will be less dra- sequences, known as sheet-like sequences of
matic, but the pressure acting on the upper sill Mount Pinafore type (after the Antarctic edifice),
surface will still be greatly reduced. The distal appear to be the products of eruptions under
part of the still-spreading magma body may thin ice (probably < 150–200 m thick; Smellie
then begin to evolve into a thicker, narrower et€al., 1993; Smellie and Skilling, 1994; Loughlin,
morphology more like that expected for a sub- 2002; Smellie, 2008). They are outflow deposits
aerial lava flow. Meltwater will be channeled (i.e., they accumulated away from the source edi-
along the side(s) of the flows, and the system fice) and are characterized by prominent basal
will remain stable as long as the pressure in beds of fluvially deposited stratified sandstones
the water above the flow(s) is maintained high rich in vitric ash clasts, for which an explosive
enough to suppress magma vesiculation to the phreatomagmatic source has been inferred.
The sequences are typically composed of (from in a shallow-marine (shelf) setting (Bergh and
base up) diamict (i.e., in situ or reworked till- Sigvaldason, 1991), were reinterpreted as prod-
ite), tephra-sourced sandstones, hyaloclastite, ucts of subglacial eruptions, possibly multiple
water-cooled lava, and, finally, subaerial lava subglacial sill intrusions and associated melt-
(Fig. 13.6). By contrast, if meltwater drainage is water floods (jökulhlaups; Smellie, 2008). The
particularly efficient beneath a thin ice cover Icelandic outcrops, which were named subgla-
(e.g., for eruptions on relatively steep bedrock), cial sheet-like sequences of Dalsheidi type, are
the vent might dry out and result in magmatic products of voluminous subglacial fissure erup-
eruptions of strombolian or hawaiian type. This tions individually up to > 30 km3. The products
was observed during the 1969 subglacial fissure can be traced up to 30 km along strike and at
eruption of basaltic andesite to andesite magma least 14 km down-dip and are separated by
at Deception Island, Antarctica, during which a sharp, undulating, likely glacial erosion sur-
line of cinder cones was generated along a fis- faces. Each sequence is formed during a single
sure in ice ~100 m thick and which were partly eruption. A “standard sequence” of four major
constructed on the ice surface (Smellie, 2002). lithofacies has been identified, i.e., diamictite,
Eruptions under much thicker ice might have lava (sill), hyaloclastite, and mudstone (from
quite different consequences for the lithofacies base up; Fig. 13.6). Although broadly resembling
formed. Unusual lava–hyaloclastite sequences sheet-like sequences of Mount Pinafore type,
of Plio-Pleistocene age at Dalsheidi in southern the Dalsheidi-type sequences are wider and usu-
Iceland, which were ascribed to lava extrusion ally much thicker (up to 300 m), and they lack
basal explosively generated phreatomagmatic with construction of a pillow ridge and simul-
tuffs and subaerial capping lava, which are sig- taneous creation of a meltwater-filled vault; (4)
nificant for indicating their formation in a sub- floating of the overlying ice and release of melt-
glacial environment (contrast Figs 13.6(a) and water in a major jökulhlaup that destabilized
13.6(b)). and destroyed much of the pre-existing pillow
In Dalsheidi-type sequences, diamictite edifice, redepositing it as thick beds of hyalo-
occurs above a basal unconformity. It prob- clastite; and (5) the high-energy flood event
ably represents a combination of tillite and locally eroding the surface of the sill. Finally, (6)
melted-out basal glacier debris, and is succeeded finer-grained beds were laid down as the flood
by laterally extensive sheet lava, called an inter- waned. The final stage was also associated with
face sill by Smellie (2008), showing evidence for intrusion of apophyses of magma derived from
basal loading and interaction with the diamic- the still-molten sill interior, which were injected
tite, which must have been relatively soft, wet, up into the sluggishly moving flow and associ-
and unconsolidated at the time. The sill shows ated hyaloclastite pile.
spectacular water-induced columnar cooling Similar to eruptions of sheet-like sequences
joints. Its upper surface is locally deformed into of Mount Pinafore type, the source edifices
structures resembling flow folds, some planed responsible for Dalsheidi-type eruptions have
off by coeval erosion, but it is also conspicuously never been observed. In part, at least, they are
characterized by prominent apophyses, usually thought to have been pillow ridges that were
a few tens of meters long, that intrude overlying extensively removed during the late-stage jökul-
massive to faintly planar stratified hyaloclastite. hlaups. Since explosive activity may have been
Thick hyaloclastite dominates each eruptive unit. absent or else confined to the final stages of
It is monomict, relatively fine grained (mainly eruptions, the sills and pillows are inferred to
coarse sand to granule grade) and fine-ash poor/ have been undegassed during emplacement due
free, with dispersed broken and intact pillows to volatile exsolution being suppressed by high
that apparently increase in proportion up-dip ambient pressures associated with unusually
toward the source. The glassy upper surface of thick ice sheet conditions (empirically calcu-
the sills has been locally stripped off by rapidly lated to be at least 1000 m). An entirely subgla-
moving, partly turbulent, sediment-charged cial setting is also suggested by the apparent
fluidal currents (known as hyperconcentrated absence of any capping subaerial lavas, indicat-
flows) that deposited the hyaloclastite as they ing that the overlying ice was never completely
slowed down. The hyperconcentrated flow melted through.
events represent periods of major meltwater It was suggested in Section 13.3.3 that a
outflows (jökulhlaups) that occurred towards combination of thicker-ice conditions (initially
the end of each eruption. The sequences are suppressing significant vesiculation) and decom-
capped by thinly stratified mudstone and fine pression caused by sudden late stage pressure
sandstone, representing finer-grained hyaloclas- reduction, e.g., as a subglacial sill nears the mar-
tite detritus deposited by a waning flood. Some gins of an ice sheet and connects with atmos-
of the finer-grained capping sediments are pheric or near-atmospheric pressures, might
formed from explosively generated lapilli tuff cause major changes within the sill. Many varia-
indicating that some eruptions were explosive tions on this theme appear to be possible. Thus
(phreatomagmatic) in the final stages. Höskuldsson et€al. (2006) described mafic pillows
A genetic model for the emplacement of inferred to have been erupted under 1.5–2.0 km
the Dalsheidi-type sequences was suggested by of ice that have outer zones with 15–20% vesicu-
Smellie (2008), comprising: (1) sill emplacement larity surrounding cores with 40–60% vesicular-
at the base of an ice sheet; (2) sill stagnation caused ity, interpreted to be the result of a ~4.5 MPa
by eruptive overpressures becoming insufficient pressure decrease as a jökulhlaup abruptly
to lift the overlying ice cover; (3) transformation removed much of the water that had been pro-
to pillow lava effusion at the erupting fissure, duced by ice melting. In contrast Schopka et€al.
(2006) described a mafic “hyaloclastite” ridge will lie on outwash in front of the glacier. This
formed under ~500 m of ice where the activity relationship raises the possibility of using these
was explosive throughout; pressure conditions features to identify the geographical limits of
were maintained low enough for magma frag- past ice sheets.
mentation by continuous drainage of meltwater
at a high enough rate that the overlying ice
could not deform fast enough to maintain con-
tact with the erupting magma. Finally, Tuffen 13.4╇ ╇ Supraglacial eruptions
(2007) has stressed the fine balance that may
exist between magma intrusion rate, water pro- Eruptions onto glacial ice may involve either the
duction and drainage rate, and ice deformation advance onto the ice of a lava flow from a vent
rate. Tuffen et€ al. (2008) described a rhyolitic located off the glacier or the deposition onto
eruption that took place beneath ~150 m of ice, the ice of pyroclasts from an explosive eruption
beginning as an explosive event but changing to (e.g., Wilson and Head, 2009). The vent for the
the intrusion of vesicular magma into the frag- latter may be located off the glacier or may lie
mental deposits from the previous explosive within the glacier in cases where a dike pen-
phase. Clearly a low-pressure pathway to the etrates into, or at least fractures, the ice lead-
atmosphere never formed in this case. ing to a phreatomagmatic eruption. To advance
The most extreme version of late-stage pres- onto a glacier, the upper surface of an encroach-
sure reduction is proposed (Section 12.3.3) to be ing lava flow must be substantially higher than
the explosive decompression of an intruded sill. the glacier so that it can override the outwardly
The loci of the explosions will migrate rapidly inclined glacier margins. Such events are prob-
back toward source (the vent), potentially gen- ably rare, though lava flow advance over thin
erating subglacial pyroclastic density currents snow and ice deposits is common. Where a
that will simultaneously mix with the abundant lower lava encounters a higher glacier surface,
ambient meltwater and rapidly transform into it will “pond” by thickening against the ice bar-
hyperconcentrated flood-flows, resulting in a rier; examples are common, both for basalts
major jökulhlaup where the floods exit from and more evolved lava flows on Earth (e.g.,
the ice sheet. There are currently no known Lescinsky and Sisson, 1998; Lescinsky and Fink,
examples of those deposits, although they 2000; Mee et€ al., 2006; Stevenson et€ al., 2006;
might easily have been missed. The deposits Harder and Russell, 2007) and Mars (Shean et€al.,
will probably be massive to weakly stratified 2005; Kadish et€ al., 2008). Another way is for
and largely formed of angular, highly vesicular lava to spill over onto a glacier from an adjacent
hyaloclasts in abundant fine ash matrix, and ice-free topographic high, although subsequent
there will be evidence for lateral (down-dip) shear through flow of the glacier will separate
and vertical transitions to fluvial deposits as the bedrock- and glacier-covering lava outcrops
floods wane and normal stream-flow condi- as the glacier moves down-valley. Examples of
tions become reestablished. Because they are truncated lavas and scoria cones are present at
generated from the destruction of the original relatively high elevations along the margins of
sill, they will not overlie any coeval sill rock, some of the Dry Valleys in Antarctica that were
which is a major distinction from “standard” formerly ice filled.
Dalsheidi-type sheet-like sequences. The depos-
its will also be composed of phreatomagmatic 13.4.1â•… Lava flowing onto glacial ice
lapilli tuffs, rather than the redeposited hyalo- Wilson and Head (2007) treated the advance
clastite (sensu White and Houghton, 2006) seen of a lava flow of a given thickness over ice
in the Dalsheidi-type sequences. Deposits that in a region where the ambient temperature
failed to escape from the ice will rest on a gla- is slightly below the ice melting tempera-
cially eroded surface and/or glacial diamict, ture. Their treatment, and that followed here,
whereas proglacial deposits of the jökulhlaup ignores the presence and likely insulating
leading to the emplacement of hot pyroclastic whereas mafic silicates have albedos as low
density currents onto glaciers, where a suffi- as 0.1 (Farrand and Singer, 1992). The corre-
cient thickness of deposit may trigger a vigorous sponding emissivities are 0.6 and 0.9, and so
interaction, especially when scouring of the ice the effect of emplacing a layer of pyroclasts
and mixing with the pyroclasts occurs, ultim- will be to increase the surface temperature
ately forming a lahar (Walder, 2000a). Chapter by a factor of up to (0.9/0.6)1/4 ≈ 1.11. Thus ice
14 discusses lahar generation and runout in at a temperature as low as (273/1.11 =) ~247 K
more detail. (−26 ºC) could be heated to the melting point
In contrast, a relatively high mass flux, high by this process. Melting will not be instantan-
volatile-content eruption (e.g., a rhyolitic plin- eous. A thermal wave will penetrate a layer of
ian eruption) will generate an eruption plume pyroclasts of thickness λ in a time τ equal to
from which the bulk of the clasts will have ~2.32 (κ τ)1/2, where κ is the thermal diffusiv-
fallen from a great enough height that they will ity, ~10–6 m2 s−1. The relevant timescales for
land at the ambient atmospheric temperature. diurnal and seasonal temperature fluctuations
If the glacier onto which they fall is stable (i.e., are one day (~9 × 104 s) and one year (~3 × 107
not in the process of melting) then they will not s), for which λ = 0.58 m and 10.6 m, respect-
immediately begin to melt it. However, pyro- ively. Thus a low-albedo pyroclast layer much
clasts are likely to have a significantly lower less than half a meter thick overlying ice in
albedo than ice, and will therefore absorb solar a region where the mean diurnal temperature
insolation more efficiently than the ice, rais- is only a few degrees below the melting tem-
ing the surface temperature of the pyroclast perature may cause a significant amount of ice
layer, possibly above the ice melting point. If melting beneath it during each daily tempera-
this temperature increase is communicated ture cycle. Conversely, a pyroclast layer a few
by conduction through the pyroclast layer to meters thick will delay the onset of ice melt-
the ice below, it will initiate ice melting. Field ing for several days, even if the surface of the
observations of tephra on glaciers suggest that pyroclast layer warms above 273 K every day.
a tephra layer <2 cm thick acts to accelerate However, a pyroclast layer more than about
ice/snow melting, while thicker layers seem to 20 m thick would be needed to protect the ice
insulate the snow surface (e.g., Manville et€ al., against an annual heating cycle.
2000, and personal observations of the authors).
However, such observations do not take time 13.4.3â•…Geological examples of
into account. On diurnal and annual times- supraglacial emplacement
cales, melting is enhanced overall, to varying Detailed descriptions of lava that flowed onto
degrees (i.e., more for thinner layers), and the glaciers are rare and focus mainly on the sec-
entire surface draped in ash is lowered at rates ondary generation of lahar and meltwater flood
above “normal” (i.e., under ash-free conditions). events (e.g., summary by Major and Newhall,
However, much thicker ash layers do insulate 1989; also Khrenov et€ al., 1988; Vinogradov
snow/ice surfaces. and Murav’ev, 1988). Smellie (2007, 2009) pos-
The warming effect can be quantified by tulated that lava emplaced on a glacier surface
considering the balance between the incom- would be redeposited immediately down-dip
ing solar heat flux and the heat flux radiated of the source edifice as breccia during rapid
by the surface, equal to (εσT4), where ε is the in situ downwasting of the glacier (e.g., at the
emissivity, σ is the Stefan-Boltzman constant, glacial termination). With coeval glacier flow,
and T is the absolute surface temperature. however, it also seems likely that most of the
Clearly, the mean surface temperature of an lava will be carried away from its source by the
exposed surface is inversely proportional to glacier and deposited as lava clast-dominated
the fourth root of its emissivity. The albedo breccia in ice-marginal moraines in much the
of glacial ice ranges up to 0.4 (Paterson, 1994), same way as cool pyroclastic deposits laid down
on moving ice. The only published description by ice flow, where they will be dumped in
of in situ lava characteristics specifically attrib- moraines and/or extensively reworked as
uted to snow interaction is by Mee et€al. (2006), further mass flows or by proglacial or later
who described an andesite lava from a winter non-glacial weathering and fluvial processes,
eruption in Chile that flowed over snow. The leaving a broad zone centered on the erupting
evidence comprised (1) a flow-front c. 5 m wide vent(s) swept free of pyroclasts (e.g., Manville
and 5 m high composed of blocky glassy brec- et€ al., 2000; Höskuldsson, 2001). The off-ice
cia, that formed as a talus apron during post- deposits will not be preserved in their original
emplacement gravitational instability; and (2) position, but will be remobilized, reworked,
a 20-m-wide zone behind the flow-front form- and redeposited. They are likely to be mixtures
ing a basal layer several metres thick, that of pyroclasts, ice, snow, and water (perhaps as
consisted of glassy andesite lava showing dis- much as 65–90% snow and ice particles in some
tinctive, cross-cutting, curviplanar “pseudop- cases; Manville et€al., 2000; Fig. 13.9). The snow/
illow” fractures and abundant perpendicular ice components will melt out, leaving behind
small-diameter secondary fractures (Fig. 13.8). a “lag” deposit comprising mainly porous
Both sets of fractures were regarded as cool- ashy material that is very susceptible to being
ing joints thought to have been caused by the washed away by rain or winnowed by wind,
overridden snow melting and flashing to steam. thus destroying any original textures. When
The lava in the snow-interaction zone was sub- mixed with blocks of accidental material, these
sequently overridden during the same emplace- lag deposits will simply resemble deposits of
ment event by crystalline subaerial lava with lahars. Of the many factors that may influence
a blocky surface autobreccia showing no evi- redistribution of pyroclasts after emplacement
dence for water chilling. on a snow and ice-covered volcano, slope angle
Pyroclasts deposited on glaciers, whether and aspect, ice thermal regime (warm-based vs.
derived from fallout or pyroclastic density cur- cold-based), pyroclast grain componentry and
rents, are reworked and redeposited by either size distribution, and a variety of local climate
eolian activity or in lahars generated by melt- parameters, such as mean temperature, diur-
ing related to pyroclastic density currents (e.g., nal temperature range, insolation, and pre-
Thouret, 1990; Walder, 2000b). They will also cipitation, are inferred to be most important
be advected en masse to marginal locations (Manville et€al., 2000).
Bergh, S. and Sigvaldason, G. E. (1991). Pleistocene Gudmundsson, M. T., Sigmundsson, F. and Björnsson,
mass-flow deposits of basaltic hyaloclastite on a H. (1997). Ice–volcano interaction of the 1996 Gjálp
shallow submarine shelf, South Iceland. Bulletin of subglacial eruption, Vatnajökull, Iceland. Nature,
Volcanology, 53, 597–611. 389, 954–957.
Björnsson, H. (1992). Jokulhlaups in Iceland: Gudmundsson, M. T., Pálsson, F., Björnsson, H. and
Prediction, characteristics and simulation Annals of Högnadóttir, T. (2002). The hyaloclastite ridge
Glaciology, 16, 95–106. formed in the subglacial 1996 eruption of Gjálp,
Blankenship, D. D., Bell, R. E., Hodge, S. M. et€al. Vatnajökull, Iceland: present day shape and future
(1993). Active volcanism beneath the West preservation. In Volcano-Ice Interaction on Earth
Antarctic Ice Sheet and implications for ice-sheet and Mars, ed. J. L. Smellie and M. G. Chapman.
stability. Nature, 361, 526–529. Geological Society of London Special Publication,
Bourgeois, O., Dauteuil, O. and Van Vliet-Lanoe, B. 202, 319–335.
(1998). Pleistocene subglacial volcanism in Iceland: Gudmundsson, M. T., Sigmundsson, F., Björnsson,
Tectonic implications. Earth and Planetary Science H. and Högnadottir, T. (2004). The 1996 eruption
Letters, 164, 165–178. at Gjálp, Vatnajökull ice cap, Iceland: efficiency of
Carslaw, H. S. and Jaeger, J. C. (1959). Conduction heat transfer, ice deformation and subglacial water
of Heat in Solids, 2nd edn. Oxford; Clarendon pressure. Bulletin of Volcanology, 66, 46–65.
Press. Harder, M. and Russell, J. K. (2007). Basanite
Corr, H. F. and Vaughan, D. G. (2008). A recent glaciovolcanism at Langorse mountain, northern
volcanic eruption beneath the West Antarctic Ice British Columbia, Canada. Bulletin of Volcanology, 69,
Sheet. Nature Geoscience, 1, 122–125. 329–340.
Dixon, J. E., Filiberto, J. R., Moore, J. G. and Hickson, Head, J. W. and Wilson, L. (1987). Lava fountain
C. J. (2002). Volatiles in basaltic glasses from a heights at Pu’u ‘O’o, Kilauea, Hawai’i: indicators
subglacial volcano in northern British Columbia of amount and variations of exsolved magma
(Canada): implication for ice sheet thickness and volatiles. Journal of Geophysical Research, 92, 13
mantle volatiles. In Volcano-Ice Interaction on Earth 715–13 719.
and Mars, ed. J. L. Smellie and M. G. Chapman. Head, J. W. and Wilson, L. (1989). Basaltic pyroclastic
Geological Society of London Special Publication, eruptions: Influence of gas-release patterns
202, 255–271. and volume fluxes on fountain structure and
Dobran, F. (2001). Volcanic Processes: Mechanisms in the formation of cinder cones, spatter cones,
Material Transport. New York, NY: Kluwer Academic/ rootless flows, lava ponds and lava flows. Journal of
Plenum. Volcanology and Geothermal Research, 37, 261–271.
Einarsson, T. (1948). The flowing lava. Studies of its Head, J. W. and Wilson, L. (2002). Mars: A review and
main physical and chemical characteristics, Societas synthesis of general environments and geological
Scientiarum Islandica, IV(3), 1–70. settings of magma-H2O interactions. In Volcano-Ice
Einarsson, T. (1994). Geology of Iceland. Rocks and Interaction on Earth and Mars, ed. J. L. Smellie and M.
Landscape. Reykjavik: Málog Menning Publishing G. Chapman. Geological Society of London Special
Company, 309 pp. Publication, 202, 27–57.
Einarsson, T. and Albertsson, K. J. (1988). The glacial Head, J. W. and Wilson, L. (2003). Deep submarine
history of Iceland during the past three million pyroclastic eruptions: theory and predicted
years. Philosophical Transactions of the Royal Society landforms and deposits. Journal of Volcanology and
London Series A, 318, 637–644. Geothermal Research, 121, 155–193.
Farrand, W. H. and Singer, R. B. (1992). Alteration Head, J. W. and Wilson, L. (2006). Heat transfer in
of hydrovolcanic basaltic ash€– observations with volcano-ice interactions on Mars: Synthesis of
visible and near-infrared spectrometry. Journal of environments and implications for processes and
Geophysical Research, 97 (B12), 17 393–17 408. landforms. Annals of Glaciology, 45, 1–13.
Geirsdóttir, A., and Eiríksson, J. (1994). Growth of Hon, K., Kauahikaua, J., Delinger, R. and
intermittent ice sheet in Iceland during the Late Mackay, K. (1994). Emplacement and inflation
Pliocene and Early Pleistocene. Quaternary Research, of pahoehoe sheet flows: Observations and
42, 115–130. measurements of active lava flows on Kilauea
Glen, J. W. (1952). Experiments on the deformation Volcano, Hawaii. Geological Society of America
of ice. Journal of Glaciology, 2, 111–114. Bulletin, 106, 351– 370.
Höskuldsson, A. (2001). Late Pleistocene subglacial Iceland€– frequency of explosive eruptions, type
caldera formation at Cero las Cumbres, eastern and volume of deposits. Journal of Quaternary Science,
Mexico. Jökull, 50, 49–64. 23(2), 109–120.
Höskuldsson, A. and Sparks, R. S. J. (1997). Lescinsky, D. T. and Sisson, T. W. (1998).
Thermodynamics and fluid dynamics of effusive Ridge-forming, ice-bounded lava flows at Mount
subglacial eruptions. Bulletin of Volcanology, 59, Rainier, Washington. Geology, 26, 351–354.
219–230. Lescinsky, D. T. and Fink, J. H. (2000). Lava and
Höskuldsson, A., Sparks, R. S. J. and Carroll, M. R. ice interactions at stratovolcanoes: Use of
(2006). Constraints on the dynamics of subglacial characteristic features to determine past glacial
basalt eruptions from geological and geochemical extents and future volcano hazards. Journal of
observations at Kverkfjöll, NE-Iceland. Bulletin of Geophysical Research, 105(B10), 23 711–23 726.
Volcanology, 68, 689–701. Lister, J. R. and Kerr, R. C. (1991). Fluid-mechanical
Jaupart, C. and Vergniolle, S. (1989). The generation models of crack propagation and their application
and collapse of a foam layer at the roof of a to magma transport in dykes. Journal of Geophysical
basaltic magma chamber. Journal of Fluid Mechanics, Research, 96, 10 049–10 077.
203, 347–380. Loughlin, S. C. (2002). Facies analysis of proximal
Jarosch, A., Gudmundsson, M. T., Högnadóttir, T. subglacial and proglacial volcaniclastic successions
and Axelsson, G. (2008). Progressive cooling of the at the Eyjafjallajokull central volcano, southern
hyaloclastite ridge at Gjálp, Iceland, 1996–2005. Iceland. In Volcano-Ice Interaction on Earth and Mars,
Journal of Volcanology and Geothermal Research, 170, ed. J. L. Smellie and M. G. Chapman. Geological
218–229. Society of London Special Publication, 202,
Jones, J. G. (1969). Intraglacial volcanoes of the 149–178.
Laugarvatn region, south-west Iceland€– I. Journal of Major, J. J. and Newhall, C. G. (1989). Snow and ice
the Geological Society of London, 124, 197–211. perturbation during historical volcanic eruptions
Jones, J. G. (1970). Intraglacial volcanoes of the and the formation of lahars and floods. Bulletin of
Laugarvatn region, south-west Iceland€– II. Journal Volcanology, 52, 1–27.
of Geology, 78, 127–140. Manville, V., Hodgson, K. A., Houghton, B. F., Keys,
Kadish, S. J., Head, J. W., Parsons, R. L. and Marchant, J. R. and White, J. D. L. (2000). Tephra, snow and
D. R. (2008). The Ascraeus Mons fan-shaped water: complex sedimentary responses at an
deposit: Volcano-ice interactions and the climatic active snow-capped stratovolcano, Ruapehu, New
implications of cold-based tropical mountain Zealand. Bulletin of Volcanology, 62, 278–293.
glaciation. Icarus, 197, 84–109, doi: 10.1016/j. Mastin, L. G. and Pollard, D. D. (1988). Surface
icarus.2008.03.019. deformation and shallow dike intrusion processes
Khrenov, A. P., Ozerov, A. Yu., Litasov, N. E. et€al. at Inyo Craters, Long Valley, California. Journal of
(1988). Parasitic eruption of Kluchevskoi volcano Geophysical Research, 93, 13 221–13 235.
(Predskazannyi eruption 1983). Volcanology and Mathews, W. H. (1947). “Tuyas,” flat-topped volcanoes
Seismology, 7, 1–24. in Northern British Columbia. American Journal of
Kieffer, S. W. (1977). Sound speed in liquid-gas Science, 245, 560–570.
mixtures: water-air and water-steam. Journal of Mee, K., Tuffen, H. and Gilbert, J. (2006).
Geophysical Research, 82, 2895–2904. Snow-contact volcanic facies and their use in
Kjartansson, G. (1948). Water flood and mud flows. In determining past eruptive environments at
The Eruption of Hekla (1947–1948). Societas Scientiarum Nevados de Chillán volcano, Chile. Bulletin of
Islandica, II(4). Volcanology, 68, 363–376.
Komatsu, G., Arzhannikov, S. G., Arzhannikova, A. Nye, J. F. (1953). The flow law of ice from
V. and K. Ershov, K. (2007). Geomorphology of measurements in glacier tunnels, laboratory
subglacial volcanoes in the Azas Plateau, the Tuva experiments, and the Jungfraufirn borehole
Republic, Russia. Geomorphology, 88, 312–328. expedition. Proceedings of the Royal Society of London
Lange, M. and Ahrens, T. (1983). The dynamic tensile Series A, 219, 477–489.
strength of ice and ice‐silicate mixtures. Journal of Parfitt, E. A. (1991). The role of rift zone storage in
Geophysical Research, 88(B2), 1197–1208. controlling the site and timing of eruptions and
Larsen, G. and Eiríksson, J. (2008). intrusions of Kilauea volcano, Hawai’i. Journal of
Late-quaternary terrestrial tephrochronology of Geophysical Research, 96, 10 101–10 112.
Parfitt, E. A., Wilson, L. and Head, J. W. (1993). discrimination. In Volcano-Ice Interaction on Earth
Basaltic magma reservoirs: factors controlling and Mars, ed. J. L. Smellie and M. G. Chapman.
their rupture characteristics and evolution. Journal Geological Society of London Special Publication,
of Volcanology and Geothermal Research, 55, 1–14. 202, 91–113.
Paterson, W. S. B. (1994). The Physics of Glaciers, 3rd. Smellie, J. L. (2000). Subglacial eruptions. In
edn. Oxford: Pergamon Press. Encyclopedia of Volcanoes, ed. H. Sigurdsson. San
Pollard, D. D. (1987). Elementary fracture mechanics Diego: Academic Press, 403–418.
applied to the structural interpretation of dikes. In Smellie, J. L. (2002). The 1969 subglacial eruption on
Mafic Dike Swarms. ed. H. C. Halls and W. F. Fahrig. Deception Island (Antarctica): events and processes
Geological Association of Canada Special Paper, during an eruption beneath a thin glacier and
34, 5–24. implications for volcanic hazards. In Volcano-Ice
Rubin, A. M. (1992). Dike-induced faulting and Interaction on Earth and Mars, ed. J. L. Smellie and M.
graben subsidence in volcanic rift zones. Journal of G. Chapman. Geological Society of London Special
Geophysical Research, 97, 1839–1858. Publication, 202, 59–79.
Rubin, A. M. (1995). Propagation of magma-filled Smellie, J. L. (2006). The relative importance of
cracks. Annual Review of Earth and Planetary Science, supraglacial versus subglacial meltwater escape in
23, 287–336. basaltic subglacial tuya eruptions: an important
Rubin, A. M. and Pollard, D. D. (1987). Origins of unresolved conundrum. Earth Science Reviews, 74,
blade-like dikes in volcanic rift zones. In Volcanism 241–268.
In Hawaii. United States Geological Survey Smellie, J. L. (2007). Quaternary vulcanism: subglacial
Professional Paper, 1350, 1449–1470. landforms. In Encyclopedia of Quaternary science, ed.
Ryan, M. P. (1987). Neutral buoyancy and the S. A. Elias. Amsterdam: Elsevier, 784–798.
mechanical evolution of magmatic systems. In Smellie, J. L. (2008). Basaltic subglacial sheet-like
Magmatic Processes: Physico-chemical Principles, ed. B. sequences: evidence for two types with
O. Mysen. Geochemical Society Special Publication, 1, different implications for the inferred
259–287. thickness of associated ice. Earth Science Reviews,
Scheu, B., Spieler, O. and Dingwell, D. B. (2006). 88, 60–88.
Dynamics of explosive volcanism at Unzen Smellie, J. L. (2009). Terrestrial sub-ice volcanism:
volcano: an experimental contribution. Bulletin of landform morphology, sequence characteristics
Volcanology, 69, 175–187. and environmental influences, and implications
Schopka, H. H., Gudmundsson, M. T. and Tuffen, for candidate Mars examples. In Preservation of
H. (2006). The formation of Helgafell, southwest Random Mega-scale Events on Mars and Earth: Influence
Iceland, a monogenetic subglacial hyaloclastite on Geologic History, eds. M. G. Chapman and L.
ridge: sedimentology, hydrology and volcano–ice Keszthelyi. Geological Society of America Special
interaction. Journal of Volcanology and Geothermal Paper, 453, pp. 55–76.
Research, 152, 359–377. Smellie, J. L. and Hole, M. J. (1997). Products and
Shean, D. E., Head, J. W. and Marchant, D. R. (2005). processes in Pliocene€– Recent, subaqueous to
Origin and evolution of a cold-based tropical emergent volcanism in the Antarctic Peninsula:
mountain glacier on Mars: The Pavonis Mons examples of englacial Surtseyan volcanic
fan-shaped deposit. Journal of Geophysical Research, construction. Bulletin of Volcanology, 58, 628–646.
110, E05001, doi:10.1029/2004JE002360. Smellie, J. and Skilling, I. P. (1994). Products of
Sigmundsson, F. and Einarsson, P. (1992). subglacial volcanic eruptions under different
Glacio-isostatic crustal movements caused by ice thickness: Two examples from Antarctica.
historical volume change of the Vatnajökull Sedimentary Geology, 91, 115–129.
ice cap, Iceland. Geophysical Research Letters, 19, Smellie, J. L., Hole, M. J. and Nell, P. A. R. (1993). Late
2123–2126. Miocene valley-confined subglacial volcanism in
Skilling, I. P. (1994). Evolution of an englacial northern Alexander Isand, Antarctic Peninsula.
volcano: Brown Bluff, Antarctica. Bulletin of Bulletin of Volcanology, 55, 273–288.
Volcanology, 56, 573–591. Stevenson, J. A., McGarvie, D. W., Smellie, J. L. and
Skilling, I. P. (2002). Basaltic pahoehoe lava-fed Gilbert, J. S. (2006). Subglacial and ice-contact
deltas: large-scale characteristics, clast generation, volcanism at Vatnafjall, Öraefajökull, Iceland.
emplacement processes and environmental Bulletin of Volcanology, 68, 737–752.
Thouret, J.-C. (1990). Effects of the November 13, Geological Society of London Special Publication,
1985 eruption on the snow pack and icecap of 202, 5–26.
Nevado del Ruiz volcano, Colombia. Journal of Wilson, L. and Head, J. W. (2007). Heat transfer in
Volcanology and Geothermal Research, 41, 177–201. volcano-ice interactions: synthesis and applications
Tuffen, H. (2007). Models of ice melting and to processes and landforms on Earth. Annals of
edifice growth at the onset of subglacial basaltic Glaciology, 45, 83–86.
eruptions. Journal of Geophysical Research, 112, Wilson, L. and Head, J. W. (2009). Tephra deposition
B03203, doi:10.1029/2006JB004523. on glaciers and ice sheets on Mars: Influence on ice
Tuffen, H. and Castro, J. (2009). The emplacement survival, debris content and flow behavior. Journal
of an obsidian dyke though thin ice: of Volcanology and Geothermal Research, 185, 290–297.
Hraftntinnuhryggur, Krafla, Iceland. Journal of Wilson, L., Fagents, S. A., Robshaw, L. E. and Scott, E.
Volcanology and Geothermal Research, (in press). D. (2007). Vent geometry and eruption conditions
Tuffen, H., McGarvie, D., Pinkerton, H., Gilbert, J. S. of the mixed rhyolite-basalt Námshraun lava
and Brooker, R. A. (2008). An explosive-intrusive flow, Iceland. Journal of Volcanology and Geothermal
subglacial rhyolite eruption at Dalakvísl, Research, 164, 127–141.
Torfajökull, Iceland. Bulletin of Volcanology, 70, Wohletz, K. H. (1986). Explosive magma-water
841–860. interactions: thermodynamics, explosive
Tuffen, H., Owen, J. and Denton, J. S. (2010). Magma mechanisms, and field studies. Bulletin of
degassing during subglacial eruptions and its use Volcanology, 48, 245–264.
to reconstruct palaeo-ice thicknesses. Earth-Science Wohletz, K. H. and McQueen, R. G. (1984).
Reviews, 99, 1–18. Experimental studies of hydromagmatic
Vinogradov, V. N. and Murav’ev, Ya. D. (1988). volcanism. In Explosive Volcanism: Inception, Evolution,
Lava-ice interaction during the 1983 Kluchevskoi and Hazards, Studies in Geophysics. Washington DC:
eruption. Volcanology and Seismology, 7, 39–61. National Academy Press, pp. 158–169.
Walder, J. S. (2000a). Pyroclast/snow interactions and Wright, A. C. (1978). A reconnaissance study of the
thermally driven slurry formation. Part 1: Theory McMurdo Volcanics north-west of Koettlitz Glacier.
for monodisperse grain beds. Bulletin of Volcanology, New Zealand Antarctic Record, 1, 10–15.
62, 105–118. Zimanowski, B., Frohlich, G. and Lorenz, V. (1991).
Walder, J. S. (2000b). Pyroclast/snow interactions Quantitative experiments on phreatomagmatic
and thermally driven slurry formation. Part explosions. Journal of Volcanology and Geothermal
2: Experiments and theoretical extension to Research, 48, 341–358.
polydisperse tephra. Bulletin of Volcanology, 62,
119–129.
Werner, R. and Schmincke, H. -U. (1999). Englacial
vs. lacustrine origin of volcanic table mountains:
Eyjafjallajökull Links
evidence from Iceland. Bulletin of Volcanology, 60,
http://en.wikipedia.org/wiki/Air_travel_disruption_
335–354.
after_the_2010_Eyjafjallaj%C3%B6kull_eruption
White, J. D. L. and Houghton, B. F. (2006). Primary
http://en.wikipedia.org/wiki/Consequences_of_the_
volcaniclastic rocks. Geology, 34, 677–680.
April_2010_Eyjafjallaj%C3%B6kull_eruption
Wilson, L. (1980). Relationships between pressure,
http://en.wikinews.org/wiki/Ash-triggered_flight_
volatile content and ejecta velocity in three types
disruptions_cost_airlines_$1.7_billion
of volcanic explosion. Journal of Volcanology and
Geothermal Research, 8, 297–313.
Wilson, L. and Head, J. W. (1981). Ascent and
eruption of basaltic magma on the Earth and Exercises
Moon. Journal of Geophysical Research. 86, 2971–3001.
Wilson, L. and Head, J. W. (2002). Heat transfer and 13.1 Why might mafic eruptions under ice be more
melting in subglacial basaltic volcanic eruptions: likely to have an explosive component than
implications for volcanic deposit morphology and subglacial rhyolitic eruptions?
meltwater volumes. In Volcano-Ice Interaction on 13.2 What is the minimum pressure required at
Earth and Mars, ed. J. L. Smellie and M. G. Chapman. the point where a sill is injected under an ice
sheet and why does this suggest that dikes Online resources available at www.cambridge.
may potentially penetrate a significant dis- org/fagents
tance into ice sheets?
• Links to websites of interest
13.3 When incandescent lava flows over ice or
• Answers to exercises
snow there is no immediate flood of melt-
water from beneath the lava. Why is this?
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K.╯P. Gregg, and Rosaly M.╯C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
subdivided into two classes based on their ini- suspension and some gravel as bedload, and
tiation: primary lahars that are caused directly have distinctive smooth, oily-appearing sur-
by eruptive activity, and secondary lahars that faces that may display polygonal convection
occur during post-eruptive or quiescent periods. cells (Fig. 14.4(a); Pierson, 2005; also see online
Both classes of lahars can occur during a sin- supplementary materials listed at end of chap-
gle eruptive sequence (e.g., Cronin et€al., 1997), ter). They exhibit density-stratified, turbulent,
or can be separated by hundreds to thousands two-phase behavior in which the solid particles
of years (Zanchetta et€ al., 2004). Lahar initi- are supported by turbulence and buoyancy, but
ation generally requires certain preconditions, as solids concentrations increase, turbulence
including: (i) an adequate water supply; (ii) becomes damped and particle–particle interac-
abundant unconsolidated sediment; (iii) gravi- tions become more important. The volumetric
tational potential; and (iv) a triggering mechan- concentration at which this transition occurs
ism. Common triggers include the mobilization strongly depends on the grain-size distribu-
of pyroclastic sediment by flowing water, due tion and clay-mineral content of the sediment
chiefly to rainfall runoff (Rodolfo and Arguden, load. Measurable static yield strengths can be
1991), eruptive expulsion of a crater lake (Suryo achieved at concentrations as low as 3 vol.% of
and Clarke, 1985; Kilgour et€ al., 2010), release smectite clays (Hampton, 1975; Fig. 14.3b) or
of an impounded water body (Björnsson, 1992; as high as 50 vol.% of coarse, neutrally buoy-
Umbal and Rodolfo, 1996; Manville and Cronin, ant particles (Bagnold, 1954; Pierson, 2005).
2007), melting of snow and ice by interaction Hyperconcentrated-flow deposits tend to be
with hot eruptive products (Major and Newhall, more poorly sorted than normal fluvial depos-
1989), and liquefaction of debris avalanches and its, but better sorted than debris-flow deposits;
debris-avalanche deposits (Scott, 1988; Stoopes they are typically weakly stratified to massively
and Sheridan, 1992; Capra et€ al., 2002; Scott textured, and are clast-supported (Fig. 14.4(b)),
et€ al., 2005). Initiation mechanism therefore reflecting particle settling through the flow and
influences the source boundary conditions (e.g., step-wise accretion at the base (Smith, 1986).
source hydrograph) used in lahar models. At solids concentrations greater than ~50%
by volume, lahars transition from hypercon-
centrated flows to debris flows. Debris flows have
volumetric solids contents of ~ 50–80%, bulk
14.3╇ ╇The sediment-concentration densities of ~ 1800–2300 kg m−3, and are com-
continuum prised of particles that range in size from clay to
boulders (Costa, 1988). Such mixtures are 104–
Lahars exhibit complex flow behavior governed 105 times more viscous than water and typically
by fluid and particle properties, and fluid–par- possess a yield strength that must be exceeded
ticle and particle–particle interactions. These before flow is possible. Once moving they can
interactions vary according to the amount (vol- achieve velocities double those of water floods
ume concentration), type, and size distribution of comparable depth and gradient due to greater
of the particles (Fig. 14.3; Pierson and Costa, bulk density, suppression of energy-dissipating
1987; Iverson, 2009). At intermediate solids con- internal turbulence, and modification of chan-
centrations (~ 20–50% by volume; bulk densities nel geometry through erosion and deposition
of 1300–1800 kg m−3), hyperconcentrated flows in order to efficiently convey flow. Interstitial
(Beverage and Culbertson, 1964; Pierson, 2005) pore fluid (water or a water-borne slurry of
have larger sediment concentrations than found fines) facilitates flow by bearing some or all of
in normal water floods. Hyperconcentrated the solids load through buoyant support and
flows typically possess a small but measurable increased fluid pressure, thereby reducing the
static yield strength, an apparent viscosity that effective stress acting on the solids and hence
is strain-rate dependent, commonly transport the intergranular friction resisting flow. Debris-
large quantities of sand-sized material in full flow mixtures commonly exhibit both dilatancy
Figure 14.3╇╇Yield strength of sediment–water mixtures as a function of suspended sediment concentration. (a) Definitions of
flow type based on an idealized yield-strength–concentration curve for a poorly sorted sediment–water mixture. (b) Measured
yield-strength–concentration curves for a range of sediment–water mixtures illustrating the effect of grain size distributions
(curves are marked with the median particle diameter) and compositions. Reprinted from Pierson (2005), with kind permission of
Springer Science+Business Media.
and contraction (where shear causes expansion non-cohesive flows (< 5% clay), and they can main-
or collapse of pore space), intergranular friction tain their high-concentration integrity for great
(due to grain–grain contacts), fluidization (due distances because they are resistant to dilution
to pore-fluid pressure exceeding hydrostatic), and transformation. Non-cohesive flows typic-
particle segregation, and minimal to moder- ally contain a narrower and coarser distribution
ate cohesion when stationary. Such flows com- of grain sizes, entrain water and deposit sedi-
monly develop a relatively coarse and dry flow ment more easily, and commonly transform dis-
front, which behaves as a moving boulder dam tally to hyperconcentrated flow (Figs. 14.5, 14.6;
(Fig. 14.4(c); Sharp and Nobles, 1953; Major and Pierson and Scott, 1985; Scott, 1988).
Iverson, 1999; Lavigne and Suwa, 2004) that can
increase maximum stage height, and a body
that usually appears to behave as a coherent but 14.4╇ ╇ Lahar characteristics
liquefied, single-phase mass or as a hypercon-
centrated flow. Debris flows commonly develop Real-time field measurements of the dynamic
steep lobate flow fronts, produce lateral levées, properties of lahars are rare, owing largely
transport cobble to boulder-sized clasts within to their size, the challenges of predicting the
the flow as well as along the channel bed, and timing of their occurrence, and financial con-
form massive, verypoorly sorted and ungraded straints in many of the countries where lahars
matrix- to clast-supported deposits (Fig. 14.4(d)). frequently occur. Nevertheless, a few field
Although debris flows can selectively deposit monitoring stations around the world have
their coarsest clasts, they can also accrete sedi- gathered data on individual events (Manville
ment incrementally (Major, 1997) or deposit it and Cronin, 2007) or on a recurrent series of
en masse as shear stresses decline and intergran- flows (Pierson, 1986; Suwa and Okuda, 1985;
ular friction, chiefly at the flow front, locks up Suwa, 1989; Ohsumi Works Office, 1995;
the flow (Major and Iverson, 1999). The presence Lavigne et€ al., 2000b), variously recording dis-
of clay-sized material can strongly influence charge hydrographs, compositions, depths,
flow behavior. Cohesive flows (defined as contain- volumes, flow-front and surface velocities,
ing > 5% clay) are typically more mobile than surface-velocity distributions, basal stresses
models lacking explicit flow physics to sophis- parameters from past lahar events and depos-
ticated physics-based models founded on con- its. They predict the average behavior of future
servation laws. Thus, lahar models fall into lahars without regard for the underlying flow
a number of principal, partially overlapping physics. Outputs from such models typically
classes. These classes include: (i) empirical mod- focus on parameters such as flow-front velocity
els based on observed correlations among lahar (i.e., arrival time), maximum discharge, stage,
parameters such as volume, flow velocity, and and inundation area.
cross-section or inundation area, but which lack
treatment of flow physics (e.g., Iverson et€ al., LAHARZ
1998; Pierson, 1998); (ii) simple rheological One of the most widely used empirical mod-
models that assume a constant stress–strain-rate els relates the area inundated by a lahar to its
relationship and composition-independent volume. On the basis of data compiled from 27
flow behavior (e.g., Fink et€ al., 1981; Manville lahars worldwide, Iverson et€al. (1998) identified
et€al., 1998); (iii) hydrologic models that assume the following relationships between lahar vol-
Newtonian behavior but are calibrated to lahars ume V (m3), channel cross-sectional area A (m2)
through modification of the flow resistance filled by a flow, and planimetric inundation area
term (e.g., Costa, 1997; Manville, 2004; Carrivick B (m2; notation is summarized in Section 14.9;
et€ al., 2009); (iv) sophisticated theoretical for- Fig. 14.9(a)):
mulations that seek to describe the constitu-
tive behavior of multi-phase mixtures (Chen, A = 0.05 V 2 / 3 , (14.1)
1988a,b; Iverson, 1997a; Iverson and Denlinger,
2001; Takahashi, 2001; Pitman and Le, 2005; B = 200 V 2 / 3 . (14.2)
Berzi et€al., 2010; George and Iverson, 2011); and
(v) mass-flow models that seek to combine elem- Solutions of Eqs. (14.1) and (14.2) were auto-
ents of approaches (iii) and (iv) with the inclu- mated in the computer algorithm LAHARZ
sion of processes such as sediment entrainment (Iverson et€ al., 1998; Schilling, 1998), a rou-
and deposition (e.g., Fagents and Baloga, 2006; tine run within the ARCINFOTM Geographical
Carrivick et€al., 2010) or development of vertical Information System (GIS). The technique was
grain-size stratification (Takahashi et€ al., 1992; designed for rapid, objective, and reprodu-
Zanuttigh and Ghilardi, 2010). cible construction of hazard maps. The model
Except for the simple empirical models, requires user-input flow volumes and a digital
lahar models generally have four principal com- elevation model (DEM) from which hydrologic
ponents: (1) a set of terms that describe conser- grids are derived for stream network identifica-
vation of mass and (sometimes) momentum of tion. Lahar deposition begins at the intersection
the bulk flow or its constituent components; (2) of a proximal hazard-zone boundary, defined
a description of channel geometry; (3) a means as the distal extent of the distance over which
of quantifying flow resistance; and (4) a means all potential sediment entrainment occurs,
of solving the resulting suite of partial differen- and a stream centerline cell (Fig. 14.9(a)), or at
tial equations numerically. The greatest differ- a user-specified stream cell. The algorithm fills
ences among the more physics-based models the cross-sectional topography at the deposi-
lie in their descriptions and treatments of flow tional starting point according to Eq. (14.1),
resistance and faithfulness to the underlying then marches downstream on a cell-by-cell basis
physics, which commonly represent a com- repeating the channel-filling process and sum-
promise between mathematical and computa- ming the planimetric area until Eq. (14.2) is satis-
tional tractability. fied. Output consists of maps delineating nested
zones of inundation potential based on the range
14.5.1â•… Empirical lahar models of user-supplied flow volumes (Fig. 14.9(b)).
Empirical models typically are founded on stat- The model has been applied widely because
istical analyses of field estimates of various of its simplicity and low data requirements
Figure 14.9╇╇ (a) Schematic illustration of the LAHARZ model methodology, modified from Iverson et€al. (1998). He represents
the drop height, and Le the horizontal distance over which potential sediment entrainment is assumed to occur before deposition
begins. (b) Example of nested lahar-inundation hazard zones predicted by LAHARZ for lahars of various volumes originating on
the flanks of San Vicente volcano, El Salvador. In this analysis, deposition begins at a user-specified location. Reprinted from Major
et€al. (2003), with permission of IOS Press.
Bingham
in Eq. (14.6); Fig. 14.11), with a viscosity η that
dilatant
(shear-thickening) (viscoplastic) is independent of shear rate. Such a material
τ = τy + ηε· m τ = τ + ηε·
y behaves as an elastic solid at shear stresses lower
Shear stress, τ (Pa)
τ = ηε·
modifies the applied normal stress; see online
supplement), produces the Coulomb-viscous
model proposed by Johnson (1970) and Johnson
Shear rate, ε· (s–1) and Rodine (1984):
Figure 14.11╇╇ Relationships between shear stress and τ = ( c + (σ − p f ) tan ϕ ) + ηε. (14.7)
shear-strain rate for common formulations of the generalized
Herschel–Bulkley rheological model. For dilatant fluids, m > 1; Viscoplastic models, however, fail to account for
for psuedoplastic fluids, m < 1. obvious particle–particle interactions inherent
in debris flows.
To account for particle collisions in debris
by rheological measurements on sub-samples flows, Takahashi (1980) developed a grain-flow
of deposits reconstituted into slurries (e.g., theory that is founded on the basis of experi-
O’Brien and Julien, 1988; Phillips and Davies, ments (Bagnold, 1954) relating normal and shear
1991; Major and Pierson, 1992; Arattano et€ al., stresses in particulate suspensions. The central
2006). The models all assume that a constant tenet of Takahashi’s theory involves a dispersive
stress–strain-rate relationship suitable for an stress τd that arises in shearing granular flows
idealized, homogeneous, single-phase material due to collisions and momentum exchanges
adequately describes debris-flow behavior. This between adjacent particles. Two regimes were
relationship is described using the generalized identified on the basis of the ratio of inertial
Herschel–Bulkley equation (Fig. 14.11): to viscous shear stresses in a suspension: a
macroviscous regime (Eq. (14.8)) dominant at
τ = τ y + ηε m , (14.6)
low shear rates; and a grain-inertia (collisional)
where τ is the applied shear stress, τy is the yield regime (Eq. (14.9)) that prevails at higher rates.
strength, η is the dynamic viscosity (i.e., the gra- These regimes are expressed as:
dient of the stress–strain-rate relationship), and τ d = λ 3 / 2ηε for ε < 40η / (λ 1/ 2 ρ s d 2 ), (14.8)
ε̇ is the shear strain rate (= ∂u/∂y). The simplest
non-Newtonian rheological model describes a τ d = ρ s (λ d)2 ε 2 for ε > 450η / (λ 1/ 2 ρ s d 2 ).
(14.9)
Bingham viscoplastic material (τy > 0 and m = 1
In these equations, ρs is the particle density, d is strength, and η is the viscoplastic viscosity. Of
a characteristic particle diameter, and λ is a lin- these parameters, h and u are calculated by the
ear grain concentration, defined as: model, whereas K, n, τy, and η are specified by
the user. The specific weight γm of a mixture
1
λ=
( )
, is obtained from its sediment concentration Cv
*
C v C v 1/ 3 − 1 and the densities of the solid and liquid phases.
(14.10)
Values of yield strength and viscosity are deter-
in which Cv* is the maximum possible static-grain mined from empirical field and laboratory obser-
volume concentration. Equation (14.9) shows vations relating Cv to τy and η (O’Brien, 2007).
that dispersive stress related to particle col- Combination of these flow-resistance terms with
lisions increases as a function of particle con- a two-dimensional hydraulic model (discussed
centration, grain size, and shear rate. Although below) enables a program (FLO-2D) that routes a
Bagnold’s results provide a tantalizing founda- hydrograph across a DEM. Expressions for mass
tion upon which to build a debris-flow theory conservation of both sediment and water, and
that incorporates effects of particle interac- a diffusive expression for momentum conser-
tions, including particle collisions that support vation (Section 14.5.4) are solved numerically.
grain-size sorting, Takahashi’s (1980) develop- FLO-2D thus assumes lahars can be described
ment of the model assumes a vertically uniform as simple, single-phase rheological materials, is
particle-concentration profile of monodisperse data intensive, and requires detailed fieldwork
particles. Those assumptions cannot explain and laboratory analyses to estimate necessary
grain-size sorting, and they place unrealistic input parameters. The model is restricted to
restrictions on development of stresses and fixed-bed (non-erodible) topography, and does
the velocity profile in a flow, and on the slope not simulate rapidly varying flow, flow surges,
angle over which material can flow (Iverson and or other flow discontinuities (such as hydraulic
Denlinger, 1987). jumps or steep flow fronts), but instead broadly
Extensions of the rheometric approach to smoothes the flow profile. Therefore, it may not
modeling lahars led to hybrid theories that be suitable for high-concentration debris flows.
combined equations in various forms. Chen Although the rheometric approach to charac-
(1988a,b) developed a hybrid approach, but one terizing and modeling debris flows (and debris-
which required multiple adjustable coefficients. flow lahars) has been widely adopted, such an
O’Brien et€al. (1993) and O’Brien (2007) summed approach has significant flaws and limitations
the viscoplastic and grain-inertial resistance (e.g., Iverson and Denlinger, 1987; Major and
terms in an energy slope form: Iverson, 1999; Hunt et€al., 2002; Iverson, 2003).
In particular, the approach assumes that debris
S f = S y + Sv + Std , (14.11) flows are rheologically simple and character-
ized by constant-valued strength and viscosity,
where Sf is the total friction slope, Sy is the yield and that particle interactions are subject to
slope, Sv is the viscous slope, and Std is the turbu- Bagnold’s dispersive-stress relations. Field and
lent-dispersive slope component. These energy- experimental measurements have shown that
slope components are written in more explicit properties of debris flows such as sediment
dimensionless form as concentration, normal stress, shear stress, and
pore-fluid pressure, all of which affect flow
τy K ηu n2u2 resistance, are spatially and temporally variable
Sf = + + 4 /3 , (14.12)
γ mh 8γ m h 2
h and that they evolve during flow. Hence, deb-
ris flows are not rheometrically simple materi-
where γm is the specific weight of the mixture, als, as properties of adjacent regions of a flow
h is flow depth, K is an empirical resistance can be quite different. Furthermore, laboratory-
parameter, u is depth-averaged flow velocity, n measured shear resistance owing to intrin-
is Manning’s roughness coefficient, τy is yield sic strength and viscosity of liquefied slurries
gradients are hydrostatic, and flow depth is using one-dimensional, unsteady streamflow
small compared to flow length. Hydraulic mod- models through manipulation of the resistance
els that solve the full Saint-Venant equations are factor, based on observations of flow behavior
referred to as fully dynamic models. and deposit geometry (Swift and Kresh, 1983;
Flow resistance is usually described by a term Laenen and Hansen, 1988; Costa, 1997; Manville,
that encompasses viscous and turbulent dissipa- 2004). Resistance factors can be derived empir-
tion, frictional losses along the channel margins, ically from simple hydraulic equations (e.g., Eq.
and changes in channel geometry. The most com- (14.16)) if local flow velocity or discharge esti-
mon resistance terms are the empirical Manning mates can be obtained and a unique relation-
coefficient n and Chézy coefficient C: ship between stage and discharge is assumed.
Discharge can be reconstructed from esti-
1 2 / 3 1/ 2 mates of flow cross-sectional area and velocity.
n= R Sf , (14.16)
u Cross-sectional area can be reconstructed from
u channel geometry, although possible scour
C= . (14.17) and deposition during flow must be carefully
RS f
assessed, and mean velocity can be estimated in
various ways, including: (i) analysis of super-ele-
In these terms, the hydraulic radius (R = A/P,
vation (tilting) of flow through a channel bend,
where P is wetted perimeter) is equivalent to
such that
flow depth for wide, shallow channels. When
combined with a description of channel geom- ke g cos θ∆e Rc
etry, an upstream boundary condition (typic- u= , (14.18)
w
ally an input hydrograph, Q(t)), an energy-loss
expression, and a downstream boundary con- where θ is the channel gradient, Δe the diffe-
dition such as a rating curve that relates h and rence in flow elevation across the channel pro-
Q, the Saint-Venant equations can be solved file, Rc the radius of curvature of the bend, w the
numerically to calculate the downstream flow width, and ke is an empirical coefficient
evolution of a flow hydrograph. Outputs typ- (Apmann, 1973); (ii) run-up against obstacles;
ically include stage, discharge, and velocity. (iii) expressions of the form u = b1h b S0b (Table
2 3
This modeling approach assumes that a lahar 14.3), where h is flow depth and S0 is channel
can be adequately characterized as a viscous bed slope; (iv) from travel-time data; or (v) from
Newtonian fluid and that flow mass and dens- independent estimates of local surface velocity
ity remain constant during flow. This is very from video-frame analysis or tracking of floating
unlikely to be the case for a highly concen- markers. Interpretation of super-elevation fea-
trated debris-flow lahar, but the approach may tures can be complicated by the non-Newtonian
provide a reasonable approximation of the behavior of sediment-laden flows and assump-
behavior of a dilute, hyperconcentrated-flow tions about the profile of the tilted flow surface;
lahar undergoing little sediment entrainment therefore, cited values of ke in Eq. (14.18) vary
or deposition. between 0.1 and 1 (Costa, 1984; Chen, 1987;
Bulmer et€al., 2002) depending on sediment con-
Chézy- and Manning-type one-dimensional centration and channel slope. Furthermore,
models estimates of mean velocity from run-up against
Although flow resistance parameters are well obstacles can be influenced by the momentum
characterized empirically for normal Newtonian of trailing flow. Discharge can also be estimated
stream flows across a range of channel morph- from stage–discharge rating curves.
ologies (e.g., Arcement and Schneider, 1989), A few studies suggest that values of Manning’s
they are less understood for high-sediment-con- n used to characterize flow resistance for lahars
centration flows approximated as Newtonian may vary with channel slope and geometry.
fluids. Nevertheless, lahars have been simulated Costa (1997) used the National Weather Service
4
1:00 hr (Qp ~ 2000 m3 s–1)
Fully dynamic flow models 2
Advances in computer processing power allow
0
fully dynamic (and multi-dimensional) flow- 1 km
routing models (i.e., models retaining all terms
in Eqs. (14.14) and (14.15)) to be applied to lahar (b)
problems. Early work (Macedonio and Pareschi, 2:00 hr
flow velocity
1992) demonstrated the use of this approach, (m s–1)
but also highlighted the importance of chan- 1 km 7
nel geometry and topographic variability, and 6
the need to adequately characterize the bound- 5
ary conditions (e.g., source hydrographs). More 4
recently, Hancox et€ al. (2001) used proprietary 3
(Mike11) software to forecast the maximum 2
discharge, stage, and arrival time at key points This inset 1:10 hr 1
0
along the flow path of the March 2007 lake-
breakout lahar from the summit Crater Lake
of Mount Ruapehu. Reconstructions of that
Figure 14.13╇╇ Sample outputs from Delft3DTM simulations
event were also made using the commercial
of the March 2007 Crater Lake break-out lahar at Mt.
Delft3D program (Fig. 14.13; Carrivick et€ al., Ruapehu, New Zealand. (a) Snapshots of flow depth at
2009). That program solves the shallow-water different time intervals showing downstream propagation of
equations for an incompressible Newtonian lahar and flooding of channel bifurcations. (b) Depth-averaged
fluid flowing over fixed channel boundar- flow velocity illustrating higher flow speeds in deeper parts
ies using the Boussinesq approximation (an of the channel. Courtesy of J. Carrivick, unpublished data. See
assumption that momentum transfer caused by color plates section.
turbulent eddies can be modeled with a user-
specified “eddy viscosity”). Reconstruction of
the Ruapehu lahar using this model reasonably 14.5.5â•… Coulomb mixture theory
matched the peak stage and inundation area of Owing to the shortcomings of modeling deb-
the flow, but poorly matched flow-front velocity ris flows using fixed rheological relationships
and time to peak stage (Manville and Cronin, between stress and strain rate, an alternative
2007). This poor replication may be due to vari- approach based on mixture theory for granu-
ous factors, including the inability of the model lar media has been developed. This approach
to treat stage-dependent roughness effects or uses separate equations for the conservation
to accommodate adequately hydraulics jumps of mass and momentum of the solid and fluid
produced by significant topographic variabil- components of a flow of granular material, and
ity, the assumption of Newtonian flow behavior coupling terms that link the momentum equa-
for what was likely a non-Newtonian flow, neg- tions of the separate phases (Hutter et€al., 1996;
lect of sediment entrainment and transport, Iverson, 1997a, 2009; Iverson and Denlinger,
and computational difficulties associated with 2001). This approach builds on descriptions of
routing a large, unsteady flow over a fine grid dry, cohesionless avalanches as depth-averaged
(Carrivick et€al., 2009). (i.e., thin-layer) granular continua governed by
Coulomb-type frictional interactions (Savage, interactions of the solid and fluid constituents in
1984; Savage and Hutter, 1989, 1991). That con- heterogeneous debris flows having high-friction,
cept was subsequently generalized to wet debris coarse-grained snouts and low-friction, lique-
flows by Iverson (1997a,b) using a simple mix- fied interiors (Iverson, 2003), (iii) the evolution
ture theory to describe the one-dimensional flow of flow depth, velocity, solids volume fraction,
dynamics of a two-phase debris flow. Since its dilatancy, and pore-fluid pressure throughout
introduction, the Coulomb mixture theory for the flow event (George and Iverson, 2011), and
debris flows has been refined to include Darcy’s (iv) post-depositional sediment consolidation
law, which allows for laminar relative motion and dissipation of non-hydrostatic pore pressure
between the fluid and solid phases, and to allow (Major and Iverson, 1999; Savage and Iverson,
coupled evolution of dilatancy, solid and fluid 2003; George and Iverson, 2011). In the limiting
volume fractions, pore-fluid pressure, flow case of dry flow, the mixture reduces to a simple
depth, and velocity (Savage and Iverson, 2003; granular Coulomb material, whereas when the
George and Iverson, 2011). Initial versions of the flow is saturated, excess pore-fluid pressure can
model were expanded from one to three dimen- balance the solids load, cause complete lique-
sions and solved using refined numerical meth- faction, and the mixture can behave as a purely
ods (Iverson and Denlinger, 2001; Denlinger and viscous fluid.
Iverson, 2001). The initial Coulomb-mixture model for deb-
Key assumptions of the Coulomb mixture ris flows (Iverson and Denlinger, 2001; Iverson,
model are: (i) the solids behave as a Coulomb 2009) consists of depth-averaged, frame-invari-
frictional material such that the shear stress ant equations that form a system of conserva-
driving flow is proportional to effective nor- tion laws for mass (Eq. (14.22)) and momentum
mal stress (σ´ = σ€ – pf ) and is independent of (Eq. (14.23)), here referred to as the debris flow
shear rate; (ii) the intergranular fluid (includ- equations (DFE):
ing fines in suspension) behaves as a Newtonian
viscous fluid; (iii) coupling between solid and ∂h ∂ ( hu x ) ∂ ( hu y )
+ + = 0, (14.22)
fluid components obeys Terzaghi’s effect- ∂t ∂x ∂y
ive stress principle (Terzaghi, 1943; Lambe
and Whitman, 1969; see online supplement)
∂ ( hu ) ∂ ( hu x2 ) ∂ ( hu x u y )
and Darcy’s law for drag owing to laminar ρ x
+ + =
relative motion of the solid and fluid phases ∂t ∂x ∂y (14.23)
(Iverson, 2003); (iv) pore-fluid pressure pf, which u2 u
reduces the intergranular effective stresses, is a − sgn ( u x ) ( ρ g z h − p bed ) 1 + x tan ϕ bed − 3v fη f x
rx g z h
‘“state”’ variable that evolves through a forced
advection-diffusion relation (Savage and Iverson,
basal shear sttresses
2003; George and Iverson, 2011), and the times- ∂ ∂p ∂2u
− hK act / pass ( ρ g z h − pbed ) − h bed + v fη f h 2x
cales for excess-pore-fluid-pressure diffusion are ∂x ∂x ∂x
long relative to the timescales of flow (Iverson, longitudinal normal stresses
1997a; Iverson and Denlinger, 2001; George and ∂u ∂ ∂2u
Iverson, 2011); (v) frictional intergranular con-
− sgn x hK act / pass ( ρ g z h − pbed ) sin ϕ + v fη f h 2x
∂y ∂y ∂y
tacts dominate shear resistance in nonliquefied transverrse shear stresses
regions of the flow (Iverson, 1997a); and (vi)
+ ρ gxh
frictional shear resistance is at least an order of
magnitude greater than viscous resistance even
gravitational body force stresses
in fully liquefied flows (Iverson, 1997a).
Under the specified assumptions, the Equation (14.23) is the x-direction (downslope)
Coulomb mixture model describes (i) the ini- momentum equation; the y-direction (trans-
tial transition of a debris flow from a static to a verse) momentum equation is obtained by
flowing state, (ii) the time- and space-dependent interchanging x and y. The direction normal to
the flow bed is noted by subscript z. Overbars runout, inundation area, velocity, flow depth,
indicate depth-averaged values. Pore-fluid and evolution of the solid volume fraction and
pressure is assumed to increase linearly with pore-fluid pressure. Predicted relaxation of
depth to a maximum pbed at the bed, and pore-fluid pressure and flow speed in the George
Kact/pass is an earth-pressure coefficient that relates and Iverson (2011) model are highly sensitive
depth-averaged lateral stresses to vertical stress to the hydraulic permeability of the mixture,
and has a value that varies with extending or a parameter that can have a value far different
compressing flow. Other terms include the fluid during flow motion than is measured in static
volume fraction vf, the pore-fluid viscosity ηf, debris.
local radius of bed curvature rx, internal φ and The Coulomb mixture-theory approach
basal φbed friction angles, downslope and normal developed by Iverson and Denlinger (2001)
components of gravitational acceleration gx, gz, was adapted to produce the simulation code
and the function sgn which changes the sign TITAN2D (Pitman et€ al., 2003; Sheridan et€ al.,
of the argument so that shear stresses always 2005). For simplicity, however, fluid terms were
oppose shear strain in the x–y plane. dropped to develop a single-phase, dry-limit
More recent advancements using the mixture model suitable for rock and dry debris ava-
theory model (George and Iverson, 2011) have lanches, and adaptive mesh refinement meth-
focused on feedbacks among and evolution of ods were adopted for computational efficiency.
mixture dilatancy, solid and fluid volume frac- A two-phase, depth-averaged formulation of
tions, and pore-fluid pressure. Feedbacks among TITAN2D was developed using engineering con-
these physical parameters affect whether the ini- cepts of mixture theory (Pitman and Le, 2005).
tial motion of a mass of granular material from a Mass and momentum conservation are calcu-
static state evolves into a rapidly moving debris lated for each phase independently using a
flow or a slowly creeping landslide, the manner Coulomb frictional model for the solids and an
and rate at which the mass moves, and the spa- assumption that the only fluid stress is pressure.
tial and temporal variations in flow depth, flow If fluid motion relative to the solid particles is
resistance, and propensity for liquefaction. not large and if pressure is the only fluid stress,
Numerical solutions of the mixture-theory Pitman and Le (2005) assume that fluid accelera-
equations have developed in parallel with the- tions can be ignored and a Darcy-type (fluid drag)
oretical advances. Solutions to early versions approximation can be used as a replacement for
of the DFE (Iverson and Denlinger, 2001) were the fluid momentum equation. Input param-
initially formulated using Riemann integration eters for TITAN2D include initial flow volume,
(Denlinger and Iverson, 2001), but were time internal and basal friction values, the ratio of
intensive. Numerical solution of the most recent solids to pore fluid, and topography. Model out-
versions of the mixture-theory equations using put includes the evolution of flow depth with
the simulation code DIGCLAW (George and time, from which various parameters such as
Iverson, 2011) adopts the latest adaptive refine- runout distance, velocity, inundation area, and
ments to solve Riemann problems (e.g., Berger discharge can be derived. Figure 14.14 shows an
et€ al., in press). Incorporation of an adaptive example of output from TITAN2D.
grid enables use of high-resolution DEMs and
increases the sensitivity of the model at the flow 14.5.6â•…Models of lahar bulking and
front. Input parameters for the mixture-theory debulking
model include initial flow volume, starting loca- Key features of lahar behavior that most mod-
tion, bed friction angle, initial solids volume els fail to address are the changes in volume,
fraction and grain density, pore-fluid density sediment loading, and downstream evolution
and viscosity, mixture hydraulic permeability that commonly occur as a result of sediment
and compressibility, coefficients that describe entrainment or deposition during transit.
mixture dilatancy, and topography in the form Changes owing to variations in sediment vol-
of a DEM. Outputs from the model include flow ume affect flow dynamics and potential areas
is specified as a function of slope); (ii) the one- an idealized homogeneous material are rela-
dimensional nature of the governing equations; tively simple to calibrate and implement, with
and (iii) inclusion of sediment entrainment and key input parameters derived from field and
loss parameters that need to be calibrated. laboratory data (Fink et€al., 1981; Whipple, 1997;
Coussot et€al., 1998; Glaze et€al., 2002; Arattano
et€al., 2006; O’Brien, 2007). However, such mod-
14.6╇ ╇ Discussion els have been extensively criticized for incom-
patibility with flow observations, ignoring
Models depicting the behavior of lahars span physical reality, use of unrealistic constant vis-
the spectrum from simple, empirical models cosity and yield strength values, and neglect
that predict the average behavior of future of dynamic particle–particle and particle–fluid
lahars on the basis of relationships among interactions that contribute to vertical and hori-
characteristics and behavior of past lahars, to zontal segregation and stratification in flows
sophisticated mathematical and numerical (Iverson, 1997a; Major and Iverson, 1999; Major,
models that describe and simulate fundamen- 2000; Iverson and Vallance, 2001; Iverson and
tal physical interactions among the constituent Denlinger, 2001; Iverson, 2003). Similarly, con-
components of flows. All models, regardless of stitutive rheological models that seek to com-
their complexity, involve compromises between bine macroviscous and collisional flow regimes
physical fidelity and computational tractability (e.g., Chen, 1988a,b ; Takahashi, 1991) have also
and efficiency. Key questions to consider when been criticized for imposing unrealistic assump-
selecting a model concern the reasons for using tions and restrictions on flow properties, and
the model, the data available, and the needs for attempting to unify physically incompatible
and expectations of those using the model out- behaviors (Iverson and Denlinger, 1987; Iverson,
put. Emergency planners and response manage- 2003). Despite those criticisms, models invoking
ment agencies are likely to have very different these approaches can replicate some features of
needs and model requirements than scientists. liquefied, high-concentration mass flows, such
During a volcanic crisis, models that empha- as plug flow, unsteady flow behavior and sur-
size simplicity, ease of use, and practical out- ging, flow depth and velocity, and deposit char-
puts such as maximum flow depth, inundation acteristics. Consequently, it can be argued that
area, and downstream arrival times will take single-phase rheological models might be useful
preference over physically faithful models that for modeling lahars reaching debris-flow con-
require extended computational timeframes, centrations if they consist of liquefied sediment
careful calibration, and an extensive set of ini- mixtures having a relatively uniform and fine
tial and boundary conditions. In some cases, grain size, particularly if they have a significant
improvements in computing power, which can clay content; in such cases, viscous flow resist-
favor use of more complex models, have been ance will dominate frictional and collisional
negated by increases in model sophistication flow resistance (Iverson, 2003). However, such
and data requirements, particularly in the form flows are likely to represent only a small frac-
of high-resolution DEMs. tion of natural lahar events.
The simplest models are heuristic, and based Shallow-water flow models draw on hydro-
on rule-of-thumb or experience-based empir- logic or hydraulic flow-routing techniques com-
ical relationships among various lahar param- mon in open-channel hydraulics, and utilize
eters. However, caution should be exercised empirical resistance terms such as Manning’s
when using an empirical model that has been coefficient or Chézy’s friction factor. Comparison
calibrated for a limited number of catchments of predicted behavior with measured behavior
or events as it may not be directly applicable of dilute hyperconcentrated-flow lahars sug-
elsewhere. gests that shallow-water models may acceptably
Rheological models that assume a constant simulate such flows (Manville, 2004; Carrivick
relationship between stress and strain rate in et€al., 2009), although these types of models are
highly sensitive to the value of the resistance triggered by lake breakouts, the forms of the
coefficient selected and flow-path topography. In source hydrographs may not be. Furthermore,
contrast to the resistance coefficients typical of the volumes and source hydrographs of lahars
normal Newtonian river floods, the Manning’s n triggered by debris avalanches, rainfall runoff,
for dilute lahars appears to increase with stage or snow-and-ice melt are more problematic to
height and sediment concentration; therefore, predict. Consequently, simulations of lahars
calibration is an issue. for purposes of assessing hazards and produ-
A more rigorous, physically faithful model cing hazard maps (as opposed to reconstruct-
based on Coulomb mixture theory (Iverson, ing events) should consider a range of initial
1997a, 2009; Iverson and Denlinger, 2001; input volumes (or hydrographs) that are guided
George and Iverson, 2011) describes the con- by analyses of past events. Dalbey et€ al. (2008)
servation of mass and momentum for the fluid discuss methods for sampling uncertain input
and solid phases separately, and specifies the parameters, propagating those uncertainties
interactions and feedbacks among mixture through a mass-flow model founded on con-
dilatancy, solid volume fraction, and pore-fluid servation laws, and quantifying the effects of
pressure which govern internal flow resistance, uncertainty on model output. Those methods
momentum transfer, and energy dissipation. allow for construction of a hazard map assess-
The model accounts for much of the observed ing the spatial probability of some parameter
heterogeneity and behavior of real debris-flow exceeding a threshold value given a range of
lahars, in which inertial and frictional effects input parameters. Error in DEMs can also affect
are significant and non-hydrostatic pore-fluid model output. Initial efforts at assessing the
pressure modulates flow behavior. However, effects of uncertainties and errors in DEMs on
it requires a number of detailed input param- geophysical mass-flow output are discussed by
eters, is computationally intensive, and its latest Stefanescu et€al. (2010).
version (George and Iverson, 2011) is presently Depending on the needs of the user, a tiered
restricted to one-dimensional numerical solu- approach for assessing lahar hazards can be
tion. Another numerical model that utilizes the adopted. For example, a rapidly produced one-
mixture-theory approach, TITAN2D, has been dimensional model, constrained by adequate
used to hindcast lahar behavior (Sheridan et€al., channel-geometry data, might be used to
2005; Williams et€al., 2008; Procter et€al., 2010). assess lahar impacts over a broad area. The out-
Simulations from that model have highlighted put from that model can then be coupled to a
its sensitivity to user-specified input parameters two- or three-dimensional model simulating
and topographic uncertainties. Furthermore, lahar impact on focused sub-reaches of critical
it requires simplifying assumptions such as interest.
neglecting the liquid phase (Pitman et€al., 2003;
Patra et€al., 2005) or imposing a Darcy-type fluid
flow that may be inappropriate for dynamic 14.7╇ ╇ Future research directions
flows or flows having very high permeability
(Pitman and Le, 2005). Significant progress has been made developing
Regardless of the model used to simulate models of lahars. For example, our understand-
lahars, output results are sensitive to uncer- ing of the physics of debris flows has advanced
tainties in input parameters, to the quality considerably, from conceptualizations of flows
and resolution of topographic models, and as single-phase materials having a constant
to propagation of uncertainties and errors stress–strain-rate relationship to more sophis-
through a model to the output. For lahars and ticated treatments of flows as granular, multi-
other geophysical mass flows, uncertainties phase materials in which the mass, momentum,
in initial volume and starting location are of and interactions of the constituent phases
first-order importance. Although initial volumes strongly influence flow dynamics. Recognition
may be reasonably well constrained for lahars that boundary conditions are extremely
important for understanding flow behavior and and Cronin, 2007; McCoy et€ al., 2010; Berger
evolution has expanded recent attention from et€ al., 2011). However, significant challenges
analyses of internal flow dynamics to interac- remain. These include capturing, in real time,
tions of flows with their boundaries. In addition the processes of vertical, lateral and longitu-
to initial efforts at modeling sediment entrain- dinal particle segregation by grain size or dens-
ment and loss by lahars (Fagents and Baloga, ity. Furthermore, a broader base of real-time
2006; Suzuki et€ al., 2009), recent experiments measurements over greater flow path distances
have examined sediment entrainment and its is needed to better resolve dynamical param-
influence on debris-flow mobility (e.g., Iverson eters and flow–boundary interactions as lahars
et€ al., 2011), and real-time field measurements evolve during transit and substrate conditions
have documented sediment entrainment by vary. Increased use of airborne and terrestrial
debris flows (McCoy et€ al., 2010; Berger et€ al., laser scanning has the potential to provide
2011). Better understanding of the conditions finer resolution of lahar–channel interactions,
that lead to sediment entrainment and loss (e.g., and the roles of sediment entrainment and
the influence of topography, channel geometry, deposition in mass conservation, than can be
and substrate composition and saturation), and obtained from discrete section measurements.
the dynamic feedbacks that lead to particle seg- Continued innovation in experimental and
regation and levée formation during flow will field techniques will refine the measurements
improve the physical fidelity of future lahar needed to fully understand the interactions
models. An ability to predict, or at least better among the constituent components of a flow,
constrain, initial lahar volumes, hydrographs, between a flow, its source characteristics and
and volumes of entrainment and loss as func- its boundary, and the mechanisms by which a
tions of time, space, and flow behavior will flow evolves as it moves downstream. Coupling
greatly improve implementation of lahar mod- of remote-sensing analyses of geodetic deform-
els and lead to more accurate assessments of ation of volcanoes, of volumes of crater lakes
downstream hazards. and potential release rates, or analyses of soil-
An increase in model sophistication and thickness distributions and landslide-stability
the use of high-resolution topography have models (such as TRIGRS; Baum et€al., 2002), with
tended to nullify simulation efficiencies gained lahar models may provide one way forward
by improved computational power. Therefore, toward predicting initial material volumes in
models adopting improved numerical tech- some cases, but predicting initial lahar volumes
niques over high-resolution, three-dimensional remains an elusive goal.
terrain are needed. Numerical finite-volume
methods that allow the computational grid to
dynamically adapt to rapidly changing condi-
tions (e.g., Berger et€al., in press; George, 2010;
14.8╇ ╇ Summary
George and Iverson, 2011) are providing steps
toward improved numerical efficiency. • Lahars are complex multiphase phenomena
As models become more sophisticated at whose behavior varies as a function of sedi-
addressing lahar physics, there is an urgent ment concentration, grain-size distribution,
need to obtain well-constrained measurements and feedbacks among the solid and fluid com-
of parameters relevant for testing those models. ponents of the mixture.
Large-scale experiments have helped derive sev- • Low-sediment-concentration lahars may
eral types of measurements needed for under- be adequately modeled using standard
standing debris-flow physics (e.g., Iverson et€al., shallow-flow techniques developed for
2010, 2011), and real-time field measurements Newtonian water flows.
of flow properties are providing important data • High-sediment-concentration lahars are best
for comparing natural events with experiments simulated using formulations based on mix-
and models (e.g., McArdell et€al., 2007; Manville ture theories that account for behaviors of,
and interactions between, the solid and fluid L runout distance (m)
constituents of a flow. Le horizontal distance from volcano
• Ranges of uncertain input parameters, espe- summit over which sediment
cially initial lahar volume, should be used entrainment is assumed in LAHARZ
when assessing downstream hazards. Methods model (m)
for sampling a range of uncertain input m exponent in Herschel–Bulkley
parameters to lahar models and for propagat- equation
ing those uncertainties through a model can Mz mean grain diameter (phi scale)
be employed to develop probabilistic hazard n Manning’s resistance coefficient
maps. (s m−1/3)
O outflow (m3 s−1)
pbed maximum pore-fluid pressure at flow
base (Pa)
14.9╇ ╇ Notation pf pore-fluid pressure (Pa)
ph hydrostatic pressure (Pa)
a exponent in regression equation pt total pressure (Pa)
a0, a1, a2 coefficients in travel-time regression p* total pressure (Pa)
equation P wetted perimeter (m)
A cross-sectional area of flow or q discharge per unit width (m2 s−1)
deposit (m2) Q discharge (m3 s−1)
b1, b2, b3 coefficient and exponents in velocity Q , Q volumetric source and sink terms
+ -
expressions (m s−1)
B planimetric area of inundation (m2) Qp peak discharge (m3 s−1)
c1, c2 coefficients in regression equations rx local radius of bed curvature (m)
c cohesive strength (Pa) R hydraulic radius (m)
C Chézy resistance term (m1/2 s−1) Rc centerline radius of curvature (m)
C m sediment mass concentration (%) s storage (m3)
Cv volumetric sediment concentration Sf friction slope
(% or fraction) S0 channel slope
Cv* maximum static sediment volume Std turbulent-dispersive slope
concentration Sv viscous slope
d particle diameter (m) Sy yield slope
Δe elevation difference across flow (m) t time (s)
g gravitational acceleration (m s−2) T(x) travel time to a point x km
gx, gz downslope and normal components downstream (hours)
of gravitational acceleration (m s−2) u mean, depth-averaged flow velocity
h flow depth (m) (m s−1)
H elevation difference between flow ux, uy x- and y-components of
source and deposit toe (m) depth-averaged velocity (m s−1)
He elevation range on a volcano over vf fluid volume fraction
which sediment entrainment is V lahar volume (m3)
assumed in LAHARZ model (m) w channel width (m)
I inflow (m3 s−1) x, y, z longitudinal, lateral, and vertical
k exponent in flux equation spatial coordinates (m)
ke coefficient in superelevation α(x) coefficient in flux equation
equation ε̇ strain rate (s−1)
K empirical resistance parameter for φ internal friction angle (˚)
viscous flow φbed bed friction angle (˚)
Kact/pass Rankine earth pressure coefficient ϕ porosity
Chen, C.╯L. (1988a). Generalized visco-plastic Cronin, S.╯J., Lecointre, J.╯A., Palmer, A.╯S. and Neall,
modelling of debris flow. Journal of Hydraulic V.╯E. (2000). Transformation, internal stratification,
Engineering, 114, 237–258. and depositional processes within a channelised,
Chen, C.╯L. (1988b). General solutions for viscoplastic multi-peaked lahar flow. New Zealand Journal of
debris flow. Journal of Hydraulic Engineering, 114, Geology and Geophysics, 43, 117–128.
259–282. Cui, P., Chen, X., Wang, Y., Hu, K. and Li, Y. (2005).
Christen, M., Kowalski, J. and Bartelt, P. (2010). Jiangjia Ravine debris flows in southwestern
RAMMS: Numerical simulation of dense snow China. In Debris-Flow Hazards and Related Phenomena,
avalanches in three-dimensional terrain. Cold ed. M. Jakob and O. Hungr. Berlin: Springer-Praxis,
Regions Science and Technology, 63, 1–14, doi: pp. 565–594.
10.1016/j.coldregions.2010.04.005 Dalbey, K., Patra, A.╯K., Pitman, E.╯B., Bursik, M.╯I.
Cole, S.╯E., Cronin, S.╯J., Sherburn, S. and Manville, and Sheridan, M.╯F. (2008). Input uncertainty
V. (2009). Seismic signals of snow-slurry lahars in propagation methods and hazard mapping of
motion: 25 September 2007, Mt. Ruapehu, New geophysical mass flows. Journal of Geophysical
Zealand. Geophysical Research Letters, 36, L09405, Research, 113, B05203, doi:10.1029/2006JB004471.
doi:10.1029/2009GL038030. Denlinger, R.╯P. and Iverson, R.╯M. (2001). Flow
Contreras, S.╯M. and Davies, T.╯R.╯H. (2000). of variably fluidized granular masses across
Coarse-grained debris-flows: Hysteresis and three-dimensional terrain 2: Numerical predictions
time-dependent rheology. Journal of Hydraulic and experimental tests. Journal of Geophysical
Engineering, 126, 938–941. Research, 106, 553–566.
Costa, J.╯E. (1984). Physical geomorphology of Dunne, T. and Fairchild, L.╯H. (1983). Estimation of
debris flows. In Development and Applications in flood and sedimentation hazards around Mt. St.
Geomorphology, ed. J.╯E. Costa and P.╯J. Fleischer. Helens. Shin Sabo, 36, 12–22.
Berlin: Springer, pp. 263–317. Dunne, T. and Leopold, L.╯B. (1981). Flood and
Costa, J.╯E. (1988). Rheologic, geomorphic, and sedimentation hazards in the Toutle and Cowlitz
sedimentologic differentiation of water floods, River system as a result of the Mt. St. Helens
hyperconcentrated flows and debris flows. In eruption. Federal Emergency Management Agency
Flood Geomorphology, ed. V.╯R. Baker, R.╯C. Kochel Report, Region X.
and P.╯C. Patton. New York: John Wiley and Sons, Fagents, S.╯A. and Baloga, S.╯M. (2006). Toward a
pp.€113–122. model for the bulking and debulking of lahars.
Costa, J.╯E. (1997). Hydraulic modeling for lahar Journal of Geophysical Research, 111(B10201),
hazards at Cascades volcanoes. Environmental and doi:10.1029/2005JB003986.
Engineering Geoscience, 3(1), 21–30. Fairchild, L.╯H. (1986). Quantitative analysis of lahar
Coussot, P. and Meunier, M. (1996). Recognition, hazard. In Mount St. Helens: 5 years later, ed. S.╯A.╯C.
classification and mechanical description of debris Keller. Washington D.C.: Washington University
flows. Earth Science Reviews, 40, 209–227. Press, pp. 61–67.
Coussot, P., Laigle, D., Arattano, M., Deganutti, A. Fink, J.╯H., Malin, M.╯C., D’alli, R.╯E. and Greeley,
and Marchi, L. (1998). Direct determination of R. (1981). Rheological properties of mudflows
rheological characteristics of debris flow. Journal associated with the Spring 1980 eruptions of
of Hydraulic Engineering, American Society of Civil Mount St. Helens volcano, Washington. Geophysical
Engineers, 124, 865–868. Research Letters, 8, 43–46.
Cronin, S.╯J., Neall, V.╯E., Lecointre, J.╯A. and Palmer, Forterre, Y. and Pouliquen, O. (2008). Flows of dense
A.╯S. (1997). Changes in Whangaehu River lahar granular media. Annual Reviews of Fluid Mechanics,
characteristics during the 1995 eruption sequence, 40, 1–24.
Ruapehu volcano, New Zealand. Journal of Fread, D.╯L. (1996). Dam-breach floods. In Hydrology
Volcanology and Geothermal Research, 76, 47–61. of Disasters, ed. V.╯P. Singh. Dordrecht: Kluwer,
Cronin, S.╯J., Neall, V.╯E., Lecointre, J.╯A. and Palmer, 85–126.
A.╯S. (1999). Dynamic interactions between lahars George, D.╯L. (2010). Adaptive finite volume
and stream flow: A case study from Ruapehu methods with well-balanced Riemann solvers for
volcano, New Zealand. Geological Society of America modeling floods in rugged terrain€– application
Bulletin, 111(1), 28–38. to the Malpasset dam-break flood (France, 1959).
International Journal of Numerical Methods in Fluids, Hutter, K., Svendsen, B. and Rickenmann, D. (1996).
66, 1000–1018 doi:10.1020/fld.2298. Debris flow modeling: A review. Continuum
George, D.╯L. and Iverson, R.╯M. (2011). A two-phase Mechanics and Thermodynamics, 8, 1–35.
debris-flow model that includes coupled evolution Iverson, R.╯M. (1997a). The physics of debris flows.
of volume fractions, granular dilatancy, and Reviews of Geophysics, 35, 245–296.
pore-fluid pressure. In Proceedings of the 5th Iverson, R.╯M. (1997b). Hydraulic modelling of
International Conference on Debris Flow Hazards unsteady debris-flow surges with solid-fluid
Mitigation, 14–17 June, 2011, Padova, Italy, ed. R. interactions. In Debris-Flow Hazards Mitigation:
Genevois, D.╯L. Hamilton and A. Prestininzi, Mechanics, Prediction, and Assessment, ed. C.-L. Chen.
Italian Journal of Engineering Geology and Environment, San Francisco, American Society of Civil Engineers,
415–424, doi:10.4908/IJEGE.2011–03.b-047. pp. 550–560
Glaze, L.╯S., Baloga, S.╯M. and Barnouin-Jha, O.╯S. Iverson, R.╯M. (2003). The debris-flow rheology
(2002). Rheologic inferences from high-water myth. In Debris-Flow Hazards Mitigation: Mechanics,
marks of debris flows. Geophysical Research Letters, Prediction, and Assessment, ed. D. Rickenmann and
29(8), 49.41–49.44. C.-L. Chen. Rotterdam: Millpress, pp. 303–314.
Griswold, J.╯P. and Iverson, R.╯M. (2008). Mobility Iverson, R.╯M. (2009). Elements of an improved
statistics and automated hazard mapping for model of debris-flow motion. Proceedings of the 6th
debris flows and rock avalanches. United States International Conference on Micromechanics of Granular
Geological Survey Scientific Investigations Report Media, 9–16.
2007–5276. Iverson, R.╯M. and Denlinger, R.╯P. (1987). The physics
Hampton, M.╯A. (1975). Competence of fine-grained of debris flows€– a conceptual assessment. In
debris flows. Journal of Sedimentary Petrology, 45, Erosion and Sedimentation in the Pacific Rim, ed. R.╯L.
834–844. Beschta, T. Blinn, G.╯E. Grant, G.╯G. Ice, and F.╯J.
Hancox, G.╯T., Keys, H. and Webby, M.╯G. (2001). Swanson. International Association of Hydrological
Assessment and mitigation of dam-break Sciences Publication 165, pp. 155–165.
lahar hazards from Mt Ruapehu Crater Lake Iverson, R.╯M. and Denlinger, R.╯P. (2001). Flow
following the 1995–96 eruptions. In Engineering of variably fluidized granular masses across
and Development in Hazardous Terrains, ed. K.╯J. three-dimensional terrain 1: Coulomb mixture
Mcmanus. Christchurch: New Zealand Institute of theory. Journal of Geophysical Research, 106, 537–552.
Professional Engineers. Iverson, R.╯M. and Vallance, J.╯W. (2001). New views
Hubbard, B.╯E., Sheridan, M.╯F., Carrasco-Núñez, of granular mass flows. Geology, 29, 115–118.
G., Díaz-Castellón, R. and Rodriguez, S.╯R. Iverson, R.╯M., Schilling, S.╯P. and Vallance, J.╯W.
(2007). Comparative lahar hazard mapping at (1998). Objective delineation of lahar-inundation
Volcan Citlaltépetl, Mexico using SRTM, ASTER hazard zones. Geological Society of America bulletin,
and DTED-1 digital topographic data. Journal of 110, 972–984.
Volcanology and Geothermal Research, 160, 99–124. Iverson, R.╯M., Logan, M. and Denlinger, R.╯P.
Huggel, C., Schneider, D., Julio Miranda, P., Delgado, (2004). Granular avalanches across irregular
H. and Kääb, A. (2008). Evaluation of ASTER and three-dimensional terrain: 2. Experimental
SRTM DEM for lahar modeling: A case study on tests. Journal of Geophysical Research, 109, F01015,
lahars from Popocatépetl Volcano, Mexico. Journal doi:10.1029/2003JF000084.
of Volcanology and Geothermal Research, 170, 99–110. Iverson, R.╯M., Logan, M., Lahusen, R.╯G. and Berti,
Hungr, O. (1990). Mobility of rock avalanches. Report M. (2010). The perfect debris flow? Aggregated
of the National Research Institute for Earth Science and results from 28 large-scale experiments.
Disaster Prevention, Japan, 46, 11–20. Journal of Geophysical Research, 115, F03005,
Hunt, M.╯L., Zenit, R., Campbell, C.╯S. and Brennen, doi:10.1029/2009JF001514.
C.╯E. (2002), Revisiting the 1954 suspension Iverson, R.╯M., Reid, M.╯E., Logan, M. et€al. (2011).
experiments of R.╯A. Bagnold. Journal of Fluid Positive feedback and momentum growth during
Mechanics, 452, 1–24. debris-flow entrainment of wet bed sediment.
Hürlimann, M., Rickenmann, D., Medina, V. and Nature Geoscience, 4, 116–121.
Bateman, A. (2008). Evaluation of approaches Johnson, A.╯M. and Rahn, P.╯H. (1970). Mobilization
to calculate debris-flow parameters for hazard of debris flows. Zeitschrift für Geomorphologie
assessment. Engineering Geology, 102, 152–163. Supplementband, 9, 168–186.
Johnson, A.╯M. and Rodine, J.╯R. (1984). Debris flow. friction concentrated at flow margins. Geological
In Slope Instability, ed. D. Brunsden and D.╯B. Prior. Society of America Bulletin, 111, 1424–1434.
Wiley, pp. 257–361. Major, J.╯J. and Newhall, C.╯G. (1989). Snow and ice
Kang, Z. and Zhang, S. (1980). A preliminary analysis perturbation during historical volcanic eruptions
of the characteristics of debris flow. In Proceedings and the formation of lahars and floods€– a global
of the International Symposium of Rover Sedimentation. review. Bulletin of Volcanology, 52, 1–27.
Beijing: Chinese Society for Hydraulic Engineering, Major, J.╯J. and Pierson, T.╯C. (1992). Debris flow
pp. 225–226. rheology: experimental analysis of fine-grained
Kaitna, R., Rickenmann, D. and Schatzmann, M. slurries. Water Resources Research, 28(3), 841–857.
(2007). Experimental study on rheologic behaviour Major, J.╯J., Schilling, S.╯P., Pullinger, C.╯R. and
of debris flow material. Acta Geotechnica, 2, 71–85. Escobar, C.╯D. (2004). Debris-flow hazards at San
Kilgour, G., Manville, V., Della Pasqua, F. et€al. (2010). Salvador, San Vicente, and San Miguel Volcanoes,
The 25 September 2007 eruption of Mt. Ruapehu, El Salvador. In Natural Hazards in El Salvador, ed.
New Zealand: Directed ballistics, Surtseyan jets, W.╯I. Rose, J.╯J. Bommer, D.╯L. Lopez, M.╯J. Carr and
and ice-slurry lahars. Journal of Volcanology and J.╯J. Major, J.J., Geological Society of America Special
Geothermal Research, 191, 1–14. Paper 375, 89–108.
Kumagai, H., Palacios, P., Maeda, T., Barba Castillo, D. Major, J.╯J., Pierson, T.╯C. and Scott, K.╯M. (2005).
and Nakano, M. (2009). Seismic tracking of lahars Debris flows at Mount St. Helens, Washington,
using tremor signals. Journal of Volcanology and USA. In Debris-flow Hazards and Related Phenomena,
Geothermal Research, 183, 112–121. ed. M. Jakob and O. Hungr. Chichester, UK:
Laenen, A. and Hansen, R.╯P. (1988). Simulation of Three Springer-Praxis, pp. 685–731.
Lahars in the Mount St. Helens Area, Washington, Using Manville, V. (2004). Palaeohydraulic analysis of the
a One-Dimensional, Unsteady-State Streamflow Model. 1953 Tangiwai lahar: New Zealand’s worst volcanic
United States Geological Survey Water-Resources disaster. Acta Vulcanologica, XVI(1/2), 137–152.
Investigation Report. Manville, V. and Cronin, S.╯J. (2007). Break-out lahar
Lambe, T.╯W. and Whitman, R.╯V. (1969). Soil from New Zealand’s Crater Lake. Eos Transactions of
Mechanics. Wiley, 553pp. the American G eophysical Union, 88, 441–442.
Lavigne, F. and Suwa, H. (2004). Contrasts between Manville, V., Hodgson, K.╯A. and White, J.╯D.╯L. (1998).
debris flows, hyperconcentrated flows and stream Rheological properties of a remobilised-tephra
flows at a channel of Mount Semeru, East Java, lahar associated with the 1995 eruption of
Indonesia. Geomorphology, 61, 41–58. Ruapehu Volcano. New Zealand. New Zealand Journal
Lavigne, F., Thouret, J.-C., Voight, B., Suwa, H. and of Geology and Geophysics, 41, 157–164.
Sumaryono, A. (2000a). Lahars at Merapi volcano: Marchi, L., Arattano, M. and Deganutti, A.╯M.
an overview. Journal of Volcanology and Geothermal (2002). Ten years of debris-flow monitoring in the
Research, 100, 423–456. Moscardo Torrent (Italian Alps). Geomorphology, 46,
Lavigne, F., Thouret, J.-C., Voight, B. et€al. (2000b). 1–17.
Instrumental lahar monitoring at Merapi Volcano, McArdell, B.╯W., Bartelt, P. and Kowalski, J. (2007).
Central Java, Indonesia. Journal of Volcanology and Field observations of basal forces and fluid pore
Geothermal Research, 100, 457–478. pressure in a debris flow. Geophysical Research Letters,
Macedonio, G. and Pareschi, M.╯T. (1992). Numerical 34, L07406, doi:10.1029/2006GL029183.
simulation of some lahars from Mount St. Helens. McCoy, S.╯W. Kean, J.╯W., Coe, J.╯A. et€al. (2010).
Journal of Volcanology and Geothermal Research, 54, Evolution of a natural debris flow: In situ
65–80. measurements of flow dynamics, video imagery,
Major, J.╯J. (1997). Depositional processes in and terrestrial laser scanning. Geology, 38, 735–738.
large-scale debris-flow experiments. Journal of Middleton, G.╯V. and Wilcock, P.╯R. (1994). Mechanics
Geology, 105, 345–366. in the Earth and Environmental Sciences. Cambridge:
Major, J.╯J. (2000). Gravity-driven consolidation of Cambridge University Press.
granular slurries€– Implications for debris-flow Muñoz-Salinas, E., Castillo-Rodriguez, M., Manea,
deposition and deposit characteristics. Journal of V.╯C., Manea, M. and Palacios, D. (2009). Lahar flow
Sedimentary Research, 70, 64–83. simulations using LAHARZ program: Application
Major, J.╯J. and Iverson, R.╯M. (1999). Debris-flow for the Popocatépetl volcano, Mexico. Journal of
deposition: effects of pore-fluid pressure and Volcanology and Geothermal Research, 182, 13–22.
Neall, V.╯E. (1996). Hydrological disasters associated Pierson, T.╯C. and Scott, K.╯M. (1985). Downstream
with volcanoes. In Hydrology of Disasters, ed. V.╯P. dilution of a lahar: transition from debris to
Singh. Kluwer, pp. 395–425. hyperconcentrated streamflow. Water Resources
O’Brien, J.╯S. (2007). FLO-2D User’s Manual, Version Research, 21, 1511–1524.
2007.06 (http://www.flo-2d.com), 78 pp. Pierson, T.╯C., Janda, R.╯J., Thouret, J.-C. and Borrero,
O’Brien, J.╯S. and Julien, P.╯Y. (1988). Laboratory C.╯A. (1990). Perturbation and melting of snow and
analysis of mudflow properties. Journal of Hydraulic ice by the 13 November 1985 eruption of Nevado
Engineering, 114, 877–887. del Ruiz, Columbia, and consequent mobilization,
O’Brien, J.╯S., Julien, P.╯Y. and Fullerton, W.╯T. (1993). flow and deposition of lahars. Journal of Volcanology
Two-dimensional water flood and mudflow and Geothermal Research, 41, 17–66.
simulation. Journal of Hydraulic Engineering, 119, Pitman, E.╯B. and Le, L. (2005). A two-fluid model for
246–261. avalanche and debris flows. Philosophical Transactions
Ohsumi Works Office (1995). Debris Flow at of the Royal Society A, 363, 1573–1601.
Sakurajima, Japan, Kyushu Regional Construction Pitman, E.╯B., Nichita, C.╯C., Patra, A. et€al. (2003).
Bureau. Computing granular avalanches and landslides.
Patra, A.╯K., Bauer, A.╯C., Nichita, C.╯C. et€al. (2005). Physics of Fluids, 15, 3638–3646.
Parallel adaptive numerical simulation of Procter, J.╯N., Cronin, S.╯J., Fuller, I.╯C. et€al. (2010).
dry avalanches over natural terrain. Journal of Lahar hazard assessment using TITAN2D for an
Volcanology and Geothermal Research, 139(1–2), 1–21. alluvial fan with rapidly changing geomorphology:
Phillips, C.╯J. and Davies, T.╯R.╯H. (1991). Determining Whangaehu River, Mt. Ruapehu. Geomorphology,
rheological parameters of debris flow material. 116, 162–174.
Geomorphology, 4, 101–110. Rickenmann, D. (1999). Empirical relationships for
Pierson, T.╯C. (1986). Flow behavior of channelized debris flows. Natural Hazards, 19, 47–77.
debris flows, Mount St. Helens, Washington. In Rodolfo, K.╯S. and Arguden, A.╯T. (1991). Rain-lahar
Hillslope Processes, ed. A.╯D. Abrahams. Boston: Allen generation and sediment-delivery systems at Mayon
& Unwin, pp. 269–296. Volcano, Philippines. In Sedimentation in Volcanic
Pierson, T.╯C. (1995). Flow characteristics of large Settings, Society for Sedimentary Geology, 71–87.
eruption-triggered debris flows at snow-clad de Saint-Venant, A.╯J.╯C.╯B. (1871). Theory of the
volcanoes: constraints for debris-flow models. nonpermanent movement of waters with the
Journal of Volcanology and Geothermal Research, 66, application to the floods of rivers and to the
283–294. introduction of the tides within their beds. Comptes
Pierson, T.╯C. (1997). Transformation of water Rendus de l’Academie des Sciences, 73, 147–154,
flood to debris flow following the 237–240.
eruption-triggered transient-lake breakout Savage, S.╯B. (1984). The mechanics of rapid
from the crater on 19 March 1982. In Hydrologic granular flows. Advances in Applied Mechanics, 24,
Consequences of Hot-Rock/Snowpack Interactions at 289–366.
Mount St. Helens volcano, Washington, 1982–84, ed. Savage, S.╯B. and Hutter, K. (1989). The motion of
T.╯C. Pierson. United States Geological Survey a finite mass of granular material down a rough
Professional Paper 1586, 19–36. incline. Journal of Fluid Mechanics, 199, 177–215.
Pierson, T.╯C. (1998). An empirical method for Savage, S.╯B. and Hutter, K. (1991). The dynamics of
estimating travel times for wet volcanic mass avalanches of granular materials from initiation to
flows. Bulletin of Volcanology, 60, 98–109. runout, I, Analysis. Acta Mechanica, 86, 201–223.
Pierson, T.╯C. (2005). Hyperconcentrated flow€– Savage, S.╯B. and Iverson, R.╯M. (2003). Surge
transitional process between water flow and debris dynamics coupled to pore-pressure evolution
flow. In Debris-flow hazards and related phenomena, ed. in debris flows. In Debris-flow Hazards Mitigation:
M. Jakob and O. Hungr. Berlin: Praxis-Springer, pp. Mechanics, Prediction, and Assessment, ed. D.
159–202. Rickenmann and C.-L. Chen. Rotterdam: Millpress,
Pierson, T.╯C. and Costa, J.╯E. (1987). A rheologic pp. 503–514.
classification of subaerial sediment-water flows. Schilling, S.╯P. (1998). LAHARZ: GIS Programs for
In Debris Flows/Avalanches: Process, Recognition, and Automated Mapping of Lahar-Inundation Hazard Zones.
Mitigation, ed. J.╯E. Costa and G.╯F. Wieczorek : United States Geological Survey Open-File Report
Geological Society of America, pp. 1–12. 98–638.
Scott, K.╯M. (1988). Origins, behavior, and Suwa, H. and Okuda, S. (1985). Measurement of
sedimentology of lahars and lahar-runout flows debris flows in Japan. Proceedings of the Fourth
in the Toutle-Cowlitz River system. United States International Conference and Field Workshop on
Geological Survey Professional Paper, 1447-A, 1–74. Landslides, Tokyo, Japan, 391–400.
Scott, K.╯M., Vallance, J.╯W., Kerle, N. et€al. (2005). Suzuki, T., Hotta, N. and Miyamoto, K. (2009).
Catastrophic precipitation-triggered lahar at Numerical simulation method of debris
Casita volcano, Nicaragua: occurrence, bulking flow introducing the non-entrainment erosion
and transformation. Earth Surface Processes and rate equation at the transition point of
Landforms, 30, 59–79. the riverbed gradient or the channel width
Sharp, R.╯P. and Nobles, L.╯H. (1953). Mudflow of 1941 and in the area of Sabo dam. Shin Sabo Journal
at Wrightwood, southern California. Geological of the Japan Society of Erosion Control Engineering,
Society of America Bulletin, 64, 547–590. 62,14–22.
Sheridan, M.╯F., Stinton, A.╯J., Patra, A. et€al. (2005). Swift, C.╯H.╯I. and Kresh, D.╯L. (1983). Mudflow Hazards
Evaluating Titan2D mass-flow model using the Along the Toutle and Cowlitz Rivers from a Hypothetical
1963 Little Tahoma Peak avalanches, Mount Failure of Spirit Lake Blockage. United States
Rainier, Washington. Journal of Volcanology and Geological Survey Water-Resources Investigation
Geothermal Research, 139, 89–102. Report 82–4125.
Smith, G.╯A. (1986). Coarse-grained nonmarine Takahashi, T. (1980). Debris flow on prismatic open
volcaniclastic sediment: Terminology and channel. Journal of the Hydraulics Division, ASCE,
depositional processes. Geological Society of America 106(HY3), 381–398.
Bulletin, 97, 1–10. Takahashi, T. (1991). Debris Flow. IAHR Monograph
Smith, G.╯A. and Fritz, W.╯J. (1989). Volcanic Series. Rotterdam: A.╯A. Balkema.
influences on terrestrial sedimentation. Geology, Takahashi, T. (2001). Mechanics and simulation of
17, 375–376. snow avalanches, pyroclastic flows and debris
Stefanescu, E.R., Bursik, M., Dalbey, K. et€al. (2010). flows. In Particulate Gravity Currents, ed. W.╯D.
DEM uncertainty and hazard analysis using a McCaffrey, B.╯C. Kneller and J. Peakall. Oxford:
geophysical flow model. In International Congress on Blackwell, pp. 11–44.
Environmental Modelling and Software, Ottawa, Canada, Takahashi, T., Nakagawa, H., Harada, T. and
ed. D.A. Swayne, W. Yang, A.╯A. Voinov, A. Rizzoli, Yamashiki, Y. (1992). Routing debris flows with
T. Filatova, S.03.01, www.iemss.org/iemss2010/ particle segregation. Journal of Hydraulic Engineering,
index.php?n=Main.Proceedings. 118, 1490–1507.
Stevens, N.╯F., Manville, V. and Heron, D.╯W. (2002). Terzaghi, K. (1943). Theoretical Soil Mechanics. New
The sensitivity of a volcanic flow model to digital York: Wiley.
elevation model accuracy: experiments with Umbal, J.╯V. and Rodolfo, K.╯S. (1996). The 1991
digitised map contours and interferometric SAR lahars of southwestern Mount Pinatubo and
at Ruapehu and Taranaki volcanoes, New Zealand. evolution of the lahar-dammed Mapanuepe
Journal of Volcanology and Geothermal Research, 119, Lake. In Fire and Mud, Eruptions and Lahars of Mount
89–105. Pinatubo, Philippines, ed. C.╯G. Newhall and R.╯S.
Stoopes, G.╯R. and Sheridan, M.╯R. (1992). Giant Punongbayan. Seattle: University of Washington
debris avalanches from the Colima Volcanic Press, pp. 951–970.
Complex, Mexico: implications for long runout Vallance, J.╯W. (2000). Lahars. In Encyclopedia of
landslides (> 100 km) and hazard assessment. Volcanoes, ed. H. Sigurdsson, B. Houghton, S.
Geology, 20, 299–302. McNutt, H. Rymer and J. Stix. New York: Academic
Suryo, I. and Clarke, M.╯C.╯G. (1985). The occurrence Press, pp. 601–616.
and mitigation of volcanic hazards in Indonesia as Vignaux, M. and Weir, G.╯J. (1990). A general model
exemplified at the Mount Merapi, Mount Kelut and for Mt. Ruapehu lahars. Bulletin of Volcanology, 52,
Mount Galunggung volcanoes. Quarterly Journal of 381–390.
Engineering Geology, 18, 79–98. Weir, G.╯J. (1982). Kinematic wave theory for
Suwa, H. (1989). Field observations of debris flow. Ruapehu lahars. New Zealand Journal of Science, 25,
Proceedings of the Japan-China (Taipei) Joint Seminar 197–203.
on Natural Hazard Mitigation, Kyoto, Japan, Wetmore, J.╯N. and Fread, D.╯L. (1984). The NWS
343–352. Simplified Dam Break Flood Forecasting Model for
Desk-top and Hand-held Microcomputers, Federal (a)╇Fluid pressure and total stress within saturated
Emergency Management Agency, 122 pp. sediment vary with depth. Consider a
Whipple, K.╯X. (1997). Open-channel flow of Bingham one-dimensional, water-saturated lahar deposit
fluids: applications in debris-flow research. Journal in which the vertical coordinate direction, y, is
of Geology, 105, 243–262. defined positive upward. If water statically fills
Williams, R., Stinton, A.╯J. and Sheridan, M.╯F. the pore space, derive an expression for the
(2008). Evaluation of the Titan2D two-phase flow hydrostatic pressure (ph) of a column of water
model using an actual event: Case study of the extending from the deposit surface, h, to a depth
2005 Vazcún Valley lahar. Journal of Volcanology and h€– y.
Geothermal Research, 177, 760–766. (b)╇Next derive an expression for the total stress σz of
Witham, C.╯S. (2005). Volcanic disasters and a column of water-saturated sediment extending
incidents: a new database. Journal of Volcanology and from the deposit surface to the same depth.
Geothermal Research, 148, 191–233. Follow the sign convention that total stress is
Worni, R., Huggel, C., Stoffel, M. and Pulgarin, B. (in negative in compression whereas fluid pressure
press). Challenges of modeling current very large is defined positive.
lahars at Nevado del Huila, Colombia. Bulletin of (c)╇Derive an expression for total mass density of the
Volcanology. deposit in terms of water density, ρw, sediment
Yano, K. and Daido, A. (1965). Fundamental study on density, ρs, and porosity, ϕ.
mud-flow. Bulletin of the Disaster Prevention Research (d)╇Using the expressions derived for total stress and
Institute, 14, 69–83. total mass density, show that
Zanchetta, G., Sulpizio, R., Pareschi, M.╯T., Leoni,
F.╯M. and Santacroce, R. (2004). Characteristics σ z = − [ρw + (ρs − ρw )(1 − φ)] g (h − y )
of May 5–6, 1998 volcaniclastic debris flows
in the Sarno area (Campania, southern Italy): This expression shows that total stress in a
relationships to structural damage and hazard column of water-saturated debris depends on the
zonation. Journal of Volcanology and Geothermal weight of the column of overlying water plus the
Research, 133, 377–393. buoyant weight of the overlying solids.
Zanuttigh, B. and Ghilardi, P. (2010). Segregation (e)╇If the fluid pressure in a lahar exactly balances
process of water-granular mixtures released down the weight of the overlying solids, the lahar
a steep chute. Journal of Hydrology, 391, 175–187. is liquefied. For that case, show that the
total fluid pressure, pt can be expressed as
pt = ρw g (h − y ) + (ρs − ρw )(1 − φ) g (h − y ) ,or
pt = ph + p*where p* is the nonequilibrium, or
Exercises excess, fluid pressure.
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K. P. Gregg, and Rosaly M. C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
distributed over a large volume surrounding gases mixed with liquid droplets, may produce
magma conduits and reservoirs. These are called velocity contrasts that are similar to, or stronger
Volcano-Tectonic (VT) earthquakes to differ than, those associated with bubbly liquids. In
entiate them from pure tectonic earthquakes, particular, dusty gases made of micron-sized
although they are indistinguishable from the particles, or misty gases made of micron-sized
latter in their broadband spectral characteris droplets, can sustain resonance at the source
tics and failure mechanisms. over durations that far exceed those achieved
Seismic signals originating from the dynam with bubbly fluids (Kumagai and Chouet, 2000).
ics of magmatic and hydrothermal fluids typ With the increased use of portable broad
ically include Long-Period (LP) events and band seismometers in the 1990s, slower proc
tremor (Chouet, 1996a). This terminology stems esses associated with unsteady mass transport
from the appearance of these signals on the began to be observed at many volcanoes. These
short-period seismometers that have tradition types of signals, with periods extending over
ally been used in volcano monitoring. LP events the range from 2 to 100 seconds, fall under the
resemble small tectonic earthquakes in dur appellation of Very-Long-Period (VLP) seismicity.
ation but differ in their characteristic frequency They are typically attributed to fluid–rock inter
range and harmonic signature (see Section 15.4). actions, as with LP events, and may involve res
Tremor is characterized by a signal of sustained onance at much longer periods than commonly
amplitude lasting from minutes to days, and observed in LP events (Kumagai, 2006), or may
sometimes for months or even longer. In many result from longer-term inertial volume changes
instances, LP events and tremor are found to in fluid-filled conduits.
have essentially the same temporal and spectral Forces associated with very large eruptions
components (Latter, 1979; Fehler, 1983), sug may also produce signals with periods extend
gesting that a common source process, differing ing beyond 100 s. For example, mantle Rayleigh
only in duration, underlies these two types of waves with periods near 230 s observed during
events. Accordingly, LP events and tremor are the Mount Pinatubo eruption of 15 June 1991
often grouped under the common appellation were linked by Kanamori and Mori (1992) to an
LP seismicity. The periods at which LP seismicity oscillatory vertical single force applied at the
is observed typically range from 0.2 to 2 seconds surface of Mount Pinatubo, which was attrib
(Chouet, 1996a), and the characteristic oscilla uted to the acoustic coupling of atmospheric
tions of LP signals are commonly viewed as a oscillations induced by the sustained energy
result of acoustic resonance in a fluid-filled cav flux from the volcano. Kawakatsu et€ al. (2000)
ity or crack (see Chapter 16). observed signals with periods >100 s associ
It is fairly straightforward to understand ated with minor phreatic activity at Aso, which
why resonance is such a pervasive phenomenon they attributed to a slow increase in fluid pres
in volcanoes. The presence of bubbles in magma sure in a source located 1 − 1.5 km below the
and hydrothermal fluids lowers the sound speed west side of Naka-dake first crater on the cen
of these fluids, inducing a sharp contrast in vel tral cone of Aso. Signals falling in this category
ocity between the fluid and encasing solid, which were classified as Ultra-Long-Period (ULP) signals
favors the entrapment of acoustic energy in the in the terminology of Chouet (1996b). Beyond
fluid volume source region. For short-lived exci the ultra-long periods are processes associated
tation, energy losses due to elastic radiation and with mass transport over timescales of min
dissipation processes at the source are the main utes, hours, and days that are more effectively
factors affecting the duration of resonance, observed with geodetic techniques. An example
hence longer-duration signals are naturally is the 30-min scale of the injection process
enhanced in low-viscosity bubbly liquids. Other recorded by Linde et€al. (1993) during the 1993
types of gaseous fluid mixtures may be even Hekla eruption.
more efficient at sustaining source resonance. It is clear that oscillatory processes are ubi
For example, gases laden with solid particles, or quitous to magma flow, and feature a large
variety of signals over a wide range of periods, When the wavelengths of observed seismic
which require very-wide-band measurements waves are much longer than the spatial extent
for their study. The present chapter offers a of the source, the source may be approximated
brief review of the state of the art in volcano by a point source and the force system repre
seismology and addresses basic issues in the sented by the moment-tensor and single-force
quantitative interpretation of processes opera components is localized at this point. The dis
tive in active volcanic systems. Our focus deals placement field generated by the point source
specifically with the quantitative analysis of may be written as (e.g., Chouet, 1996b)
signals originating in the movement of magma
â•… u n ( x, t) = M pq (t) ∗ G np ,q ( x, ξ, t ) + Fp (t ) ∗ G np ( x, ξ, t ) ,
and/or hydrothermal fluids. Other areas in the
(15.1)
very broad discipline of volcano seismology,
such as the tomographic method used in the where un(x,t) is the n-component of seismic dis
elaboration of the three-dimensional velocity placement observed at a point x at time t, Mpq(t)
structures of volcanoes, array-processing meth is the time history of the pq-component of the
ods used in tracking tremor sources, or eruption moment tensor at position ξ of the source, Fp(t)
monitoring involving detailed investigations of is the time history of the force applied in the
seismicity, are not addressed here. Perspectives p-direction at ξ, and Gnp(x,ξ,t) is the Green tensor
on these and further aspects of the field may which relates the xn-component of displacement
be found in the reviews by Chouet (1996b, at x with the xp-component of impulsive force
2003), McNutt (2002, 2005), Konstantinou applied at ξ. The notation q indicates spatial dif
and Schlindwein (2002), and Kawakatsu and ferentiation with respect to the ξq-coordinate
Yamamoto (2007). at the source and the symbol * denotes convo
lution. The summation convention is assumed
throughout for repeated subscripts. All notation
is given in Section 15.7.
15.2╇ ╇Description of seismic In Eq. (15.1), the derivative of Gnp with
sources in volcanoes respect to the source coordinate ξq represents
the contribution from a single couple with arm
A general kinematic description of seismic in the ξq-direction at ξ. The strength of the p − q
sources in volcanoes is commonly based on a couple is given by Mpq, which has dimensions of
moment tensor and single force representa moment (Nm). The sum over q implied by the
tion of the source (Aki and Richards, 2002). The repeated indices states that each displacement
seismic moment tensor allows a description component at x is made of a sum of couples
of any generally oriented discontinuity in the located at ξ. With three components of force
Earth (such as slip across a fracture plane, or and three possible arm directions, nine force
the opening of a crack) in terms of equivalent couples can be defined at ξ, with each corre
body forces. As mass-advection processes can sponding to one set of opposing forces (dipoles
also generate forces on the Earth, a complete or shear couples) (Fig. 15.1). The combination
description of volcanic sources further requires of the nine couples Mpq characterizes all the
the consideration of single forces in addition to temporal information about a fault or conduit
the geometrical source components described that can be extracted from the observation of
by the moment tensor. The standard approach waves whose wavelengths are much larger than
to estimate the source-time histories of the the spatial extent of the source. The compo
moment and force components at the source is nents of the moment tensor describing a fault
based on the Green’s function which describes or conduit embedded in an isotropic medium
the signal propagating through a known vel are given by the expression (Aki and Richards,
ocity structure that would be observed at a 2002)
receiver if the source-time function were a per
fect impulse. M pq = ∫∫ [ u i ]ν j [ λδ ij δ pq + µ (δ ip δ jq + δ iq δ jp )]dΣ . (15.2)
Σ
In this equation, Σ represents the surface of the slip [u1] parallel to Σ lying in the plane ξ3 = 0 (Fig.
buried fault or conduit across which the dis 15.2 a). In this case, ν has components (0,0,1), [u]
placements are discontinuous, νj is the j-th com has components ([u1],0,0), and the moment ten
ponent of the unit vector normal to the surface sor is obtained as
element dΣ, [ui] denotes the i-th component of 0 0 1
the displacement discontinuity between the Σ+
M = M (t) 0 0 0 , (15.3)
side of Σ and the Σ− side of Σ, with square brack
1 0 0
ets referring to the difference u|Σ+ − u|Σ−, λ and
μ are the Lamé parameters (elastic moduli) of
the rock matrix, and δij is the Kronecker symbol where
(δij = 0 for i ≠ j, and δij = 1 for i = j). The seismic
M (t) = µ ∫∫ [ u1 (ξ, t)]dΣ (15.4)
moment tensor is a symmetric second-order Σ
tensor (Mpq = Mqp) and thus has six independent represents the moment source-time function.
components. Examples of moment tensors for Equation (15.3) is the familiar double couple
fault slip and volumetric sources are discussed representing a slip source (Fig. 15.2 b). The static
below. seismic moment for the effective point source
of slip is given by
15.2.1â•…Seismic moment tensor for M0 = µ[ u1 ] S , (15.5)
representative sources
To illustrate fault slip associated with the shear where [ u1 ] is final value of slip averaged over the
failure of rock, let us consider a Cartesian coord fault surface (∫∫ [ u1 (ξ, t → ∞)]dΣ / S) and S is the
Σ
inate system (ξ1,ξ2,ξ3) at the source and assume a fault area. The product area × slip characterizes
(a) ξ3
1 0 0
M′ = M (t) 0 0 0 . (15.6)
0 0 −1
ξ1 ξ2 (15.8)
(a) (b)
ξ3 ξ3 ξ3� Figure 15.3╇╇ (a) Tensile crack
ν opening in the direction ν defined by
the angles θ and ϕ.
θ (b) Representation of the crack as
θ
three vector dipoles, in which each
dipole consists of a pair of forces
offset in the direction of the force.
ξ2�
ξ2 ξ2
ξ1 φ ξ1 φ ξ1�
ξ2�
ξ2 ξ2
ξ1 ξ1
φ φ
ξ1�
(Chouet et€ al., 2003, 2005; Waite et€ al., 2008), axes as coordinate axes, the components of M
would have ratios of [1:1:2]. are represented by three dipoles with magni
The moment tensor corresponding to the tudes λΔV, (λ + μ)ΔV, and (λ + μ)ΔV, where the
radial expansion of a cylinder whose axis orien smallest dipole is oriented along the pipe axis
tation is given by θ and ϕ (Fig. 15.4 a) is obtained (Fig. 15.4 (b)). For λ = μ (Poisson solid), the dipole
in a similar manner as (Chouet, 1996b) magnitude ratios become [1:2:2], while for λ = 2μ
(rock near liquidus temperatures) these reduce
M = ∆V (t)
to [1:3/2:3/2].
λ + µ(cos2 θ cos2 φ + sin 2 φ −µ sin 2 θ sin φ cos φ −µ sin θ cos θ cos φ The moment tensor for a spherical source is
− µ sin θ sin φ cos φ
2
λ + µ (cos2 θ sin 2 φ + cos2 φ −µ sin θ cos θ sin φ , given by (Aki and Richards, 2002)
−µ sin θ cos θ cos φ −µ sin θ cos θ sin φ λ + µ sin θ
2
(15.9) λ + 23µ 0 0
where ΔV(t) = LΔS(t), in which L is the pipe M = ∆V (t) 0 λ + 23µ 0 , (15.10)
length and ΔS(t) is the increase in the pipe 0 0 λ + 23µ
cross-sectional area. Referred to the principal
where ΔV(t) = 43πR3ΔΘ(t), in which R is the volume changes are related through the expres
radius and ΔΘ(t) is the fractional change in vol sion (Sneddon and Lowengrub, 1969)
ume of the sphere. The spherical source has an
δ p = 34µ (λ + µ)λ + 2µδ VR 3 . (15.15)
equivalent force system made of three equiva
lent vector dipoles. Kawakatsu and Yamamoto (2007) noted that δV
When estimating the magnitude of ΔV, represents the volume change used in geodetic
a clear distinction must be made between a analyses (e.g., Mogi, 1958) and proposed to call
source that is unconstrained and free to expand this volume the “Mogi volume’’ to distinguish it
and a source confined by the pressure of the from the stress-free volume change.
surrounding medium. The volume change ΔV In general, magma movement between
obtained from the moment tensor represents adjacent segments of conduit or reservoir can
the stress-free volumetric strain introduced by be represented through a combination of volu
Eshelby (1957) (see Aki and Richards, 2002, pp. metric sources such as described above. For
53–54 for a discussion) and is strictly applicable example, let us consider the transfer of magma
to a source that is free to expand. Under con from a spherical chamber into a vertical dike.
fining pressure from the surrounding medium, Assuming a crack opening in the direction θ =
the actual volume increase may be smaller than π/2, ϕ = 0 (see Fig. 15.3 (a)), the moment ten
ΔV. That is, rather than expanding by ΔV and sor given by Eq. (15.8) is purely diagonal with
being subjected to zero pressure, the source components (λ + 2μ)ΔV, λΔV, and λΔV. Using
may expand by δV subjected to a pressure Eq. (15.10), with shorthand notation (1,1,1) to
increase δp, where the ratio δV/ΔV depends on represent the diagonal matrix in this equation,
the geometry of the reservoir. In a spherical and applying mass conservation, we obtain the
reservoir, the actual volumetric strain is related moment components in the spherical cham
to the stress-free volume increase by (Aki and ber as −(λ + (2/3)μ)ΔV(1,1,1). If the two volume
Richards, 2002) changes share the same source-time function,
the resulting source should be observed as the
δ V = λ + 23µλ + 2µ∆V , (15.11)
sum of the two sources with tensor compo
and the corresponding pressure increase is given nents (2/3)μΔV(2,−1,−1), which shows expan
by (Aki and Richards, 2002) sion in the ξ1 direction and contraction in the
ξ2 and ξ3 directions (see source coordinates in
δ p = 43µδ VV . (15.12)
Fig. 15.3 (a)). The isotropic component of the
The volume change in this case may range from moment tensor is zero and this source is known
approximately half the size of the stress-free as a compensated linear vector dipole (CLVD)
volume change when λ = μ, to two-thirds of the (Knopoff and Randall, 1970). Other geometrical
stress-free volume change when λ = 2μ. For a configurations of sources may produce a simi
cylindrical source, one obtains (Kawakatsu and lar CLVD mechanism (Chouet, 1996b). In prac
Yamamoto, 2007) tice, however, observations are limited to those
δ V = λ + µλ + 2µ∆V , portions of conduit that provide specific sites
(15.13)
where pressure and momentum changes in the
and fluid are effectively coupled to the Earth, and
δ p = µδ VV . the isotropic component of the associated diag
(15.14)
onalized moment tensor is non-zero.
For this geometry, δV may range from two-thirds
of the stress-free volume change when λ = μ, to 15.2.2â•… Single force
three-quarters of the stress-free volume when The single force in Eq. (15.1) represents the
λ = 2μ. exchange of linear momentum between the
In the case of a thin crack, the two volume source and the Earth (Takai and Kumazawa,
changes are the same (δV = ΔV). In a penny- 1994). Whenever some mass gains momen
shaped crack with radius R, the pressure and tum, the counter force due to this accelerating
FB FB
Down
0 te t 0 t
te
mass is felt by the Earth. When the mass even cylindrical cavity initially sealed at the top by
tually decelerates and comes to rest, it induces a lid (Fig. 15.5 (a)). A pressurized inviscid fluid
another force in the Earth in the same direction. in the cavity exerts an upward vertical force FT
Momentum conservation requires that the net on the lid, a horizontal outward force FS on the
change of linear momentum in the overall source sidewall, and a vertical downward force FB on
system must cancel out over the total duration the bottom of the cylinder. The eruption is sim
of an event (Takei and Kumazawa, 1994), so that ulated by the sudden removal of the lid at time t
the two pulse-like single forces associated with = te, at which point the force FT vanishes instant
the acceleration and deceleration phases must aneously and the fluid pressure in the cylinder
counterbalance each other. starts to decrease with a characteristic time con
Landslides are examples of external stant τ fixed by the mass flux of the eruption,
single-force sources (Kanamori and Given, i.e., τ ~ ℓ/v, where ℓ is the length of the cylin
1982). On a timescale of minutes a landslide der and v is the mean fluid velocity inside the
may be viewed as a box sliding down a slope. cylinder (assuming purely vertical flow and no
As the sliding mass gains momentum, a counter turbulence). Since the forces FS and FB are both
force is felt by the Earth in the up-dip direction. proportional to pressure, they decrease with the
When the sliding mass eventually loses momen same time constant τ. The time histories of FT, FS,
tum, a frictional force is applied to the under and FB are sketched in Figure 15.5 (b). The force
lying ground in the down-dip sliding direction. FT can be decomposed into a vertical downward
Kawakatsu (1989) developed a centroid single component FT1 and vertical upward component
force (CSF) inversion method and applied it to FT2 in such a way that FT2 has the same time his
analyses of landslide and slump events recorded tory as FS and FB (Fig. 15.5 (c)). As a result, the three
by the global seismic network. The CSF inver forces FT2, FS, and FB form an implosive source so
sion method was applied by Ekström et€al. (2003) that the eruption mechanism is represented by
to quantify the stick-slip, downhill sliding of a the superposition of a downward vertical force
glacial ice mass. (the reaction force of the volcanic jet) with this
Another example of external force is the volumetric implosion. In the model considered
recoil force in the equivalent force system rep here, the implosive source has a moment ten
resenting a volcanic eruption in the model of sor given by Eq. (15.9), but other geometries are
Kanamori et€al. (1984). In this model, Kanamori possible as given by Eqs. (15.8) and (15.10). The
et€al. (1984) consider a shallow vertically oriented above model was used by Uhira and Takeo (1994)
that as the slug ascends and expands, viscous which can be recast in matrix form as
shear along the tube wall provides support for U (ω) = G (ω)S(ω) ,
(15.17)
an increasing mass of liquid. Consequently, the
static pressure below the slug decreases, reducing where U is a Nt × 1 vector of Fourier-transformed
the downward force exerted by pressure on the ground displacement components, G is the Nt × 9
upward-facing base of the tube and effectively matrix of Fourier transforms of the Green’s func
imparting an upward force on the tube (see Fig. tions, S is the 9 × 1 vector of Fourier-transformed
6.6 (b)). During most of the ascent, this increas force and moment-tensor components (three
ing upward force on the tube due to decreasing force components and six independent moment
basal pressure is compensated by the increasing components), and Nt is the number of observed
downward shear force exerted on the tube wall seismic traces. Inverting for each frequency
by the descending film of liquid surrounding the component separately by minimizing the least
slug body. Downward shear gradually increases squares residuals between data and synthetics,
as the slug expands and lengthens, until rapid one obtains estimates for the source:
expansion of the slug just prior to reaching
Sest (ω) = [G H (ω) G (ω)]−1 G H (ω) U (ω) , (15.18)
the surface induces a component of upward-
directed shear in the region around and above
the slug nose (see Fig. 6.6 (a)), which opposes where the symbol H indicates the conjugate
most of the downward shear due to the descend transpose (Hermitian) and Sest is the vector of
ing liquid film around the slug body. Upon calculated Fourier-transformed source compo
slug burst, expansion no longer drives upward nents. After solving for all frequencies of inter
shear. Upward shear is then provided solely by est, one obtains the time domain estimates of
liquid inertia and declines rapidly, leaving only the force and moment-tensor components of
the downward shear force, which subsequently the source with the inverse Fourier transforms
decays slowly back to zero as the draining film of the relevant components of the vectors Sest.
thins out. The slumping of the liquid film sur Synthetic seismograms for all the traces are
rounding the slug back to the top of the liquid computed in the frequency domain using Eq.
column after the slug has burst induces a sim (15.16) followed by application of the inverse
ultaneous rise in static pressure and increase in Fourier transform.
basal pressure. The slug expansion and burst thus Working in the frequency domain reduces
result in a net upward force transient with both the computational load by allowing the inver
shear and pressure components being exerted sion of many small matrices, which is more effi
on the tube. This model was used by Chouet et€al. cient than inverting a single very large matrix
(2010) to interpret the seismic source mechan as required in the time-domain approach (Auger
ism of degassing bursts at Kilauea. et€al., 2006). It also reduces the number of sam
ples required to model the signal. For example, a
trace of 200 s sampled at 50 samples per second
yields 10 000 time samples. In the frequency
15.3╇ ╇Waveform inversion domain, the same trace represents 200 samples
per Hz. For a VLP data inversion, in which we
With knowledge of the Earth’s structure and the may typically consider a limited frequency band
associated Green’s functions, the source-time from 0.01 to 0.5 Hz, the frequency-domain inver
histories of the time-dependent moment tensor sion requires just 100 spectral components.
and force components can be retrieved through For an accurate determination of the source
least squares inversion of observed waveforms. location and associated source mechanism, one
The frequency-domain version of Eq. (15.1) is needs to compare observed seismic data to syn
thetic data calculated for a realistic model of
u ( x, ω) = M (ω) ⋅ G ( x, ξ, ω) + F (ω) ⋅ G ( x, ξ, ω) , the volcanic edifice. A standard approach is to
n pq np , q p np
tensor. Rather, the time histories of individual the summit crater, and the source-time his
moment-tensor components are obtained inde tory features two downward force components
pendently of each other (e.g., Figs. 15.8 (a) and separated by an upward force component. The
15.8 (b)). For a realistic interpretation of the initial downward force component was attrib
source mechanism consistent waveform shapes uted by Ohminato et€ al. (2006) to the sudden
among individual moment tensor components removal of the lid capping the pressurized con
are required. Differences among the time his duit (this force is analogous to the force FT1 in
tories of individual moment tensor components Fig. 15.5 (c)), and the subsequent upward force
may arise due to the presence of noise in the was interpreted by these authors as a drag force
data, inadequate receiver coverage, or an inad induced by viscous magma moving up the con
equate starting assumption concerning the duit. Ohminato et€al. (2006) attributed the final
source mechanism. In the latter case, the intro downward force component to an explosive
duction of single force components may help fragmentation of the magma, whose effect is to
minimize distortion of the source-time func effectively cancel viscous drag so that the down
tions of moment components (e.g., Chouet et€al., ward force due to jet recoil again dominates the
2003). For a robust estimation of the source signal. An application of the waveform inver
mechanism, appropriate network coverage of sion method to VLP data recorded at Stromboli
the entire volcanic edifice is required. Ideally, Volcano, Italy, is presented below.
ten or more three-component receivers ring
ing the edifice within a range of 5 km from the 15.3.1â•…Shallow conduit dynamics at
source should be considered. At the very least, Stromboli imaged from VLP data
five three-component receivers surrounding Detailed broadband measurements were car
the source at close range are required to gain a ried out at Stromboli in September 1997 by
rough idea of the source mechanism, provided Chouet et€al. (2003) using a network of 21 three-
noise in the data is not an issue (Dawson et€al., Â�component broadband (0.02–60 s) seismometers
2011). The moment tensor estimated from wave (Fig. 15.7). Eruptive activity at that time was lim
form inversion may be interpreted through a ited to two vents located at the northern and
comparison with the theoretical moment ten southern perimeters of the crater, and two char
sors given by Eqs. (15.8), (15.9), and (15.10), or acteristic types of waveforms representative
geometries made of composites of such sources of eruptions from these vents were observed.
(see example in Section 15.3.1 below). The signals associated with eruptions from the
Inversions of VLP waveforms have imaged northern vent were subsequently named Type-1
crack geometries at the source in the form of events, and those related to eruptions from the
dikes or sills (Ohminato and Chouet, 1997; southern vent were named Type-2 events. Using
Chouet et€al., 2003; Kumagai et€al., 2003), as well these data, Chouet et€al. (2003) carried out sys
as more complicated geometrical configurations tematic inversions of eruption signals band-pass
involving a composite of a dike intersecting a sill filtered in the 2–20 s band for a Type-1 event,
(Chouet et€al., 2005), composites of intersecting or 2–30 s band for a Type-2 event. Two best-fit
dikes (Chouet et€ al., 2008, 2010), or two cham models, based on residual error E (Eq. (15.20))
bers linked to each other by a narrow channel and using six moment-tensor components plus
(Nishimura et€al., 2000). Contrasting with these three single-force components, were imaged for
findings are results obtained by Ohminato et€al. two point sources located at elevations of 520
(2006) from waveform inversions of VLP signals m (Type-1) and 480 m (Type-2) above sea level,
produced by vulcanian explosions at Asama. At approximately 160 m northwest of the vents
Asama, the contribution from a vertical single (Chouet et€al., 2003).
force with magnitude 1010 − 1011 N was found to Figures 15.8 (a) and 15.8 (b) show the source-
dominate the observed waveforms, and no obvi time functions of moment and force compo
ous volumetric component was identified at the nents obtained by Chouet et€al. (2003) for Type-1
source. The depth of the source is ∼200 m below and Type-2 events, respectively. The consistency
of waveform shapes seen among individual Accompanying the volumetric source com
moment tensor components points to a robust ponents in Figures 15.8 (a) and 15.8 (b) is a dom
underlying mechanism. The principal axes of inantly vertical single force. The force is initially
the moment tensor obtained by eigenvalue down, then up in both event types. In the Type-1
decomposition of the source-time functions event, the downward force is synchronous with
of the moment components are illustrated in the initial inflation of the source volume while
Figures 15.8 (c) and 15.8 (d). The three eigenvec the following upward force coincides with a
tors identified by the bold black arrows in these deflation of the source volume. Although less
figures are obtained from measurements of the clear, a similar synchronicity is manifest in the
maximum peak-to-trough amplitudes in the Type-2 event. Taken together, this force and vol
individual tensor components. ume change may be viewed as the result of a
The force system in Figures 15.8 (c) and piston-like action of the liquid in response to
15.8 (d) consists of three dipoles with ampli the disruption of a gas slug transiting through
tude ratios [1:0.8:2] and [1.1:1:2] in the Type-1 this particular location in the conduit (James
and Type-2 events, respectively. These ratios et€al., 2006).
closely match the amplitude ratios [1:1:(λ + As noted above, the amplitude ratios of the
2μ)/λ] for a crack, in which λ = 2μ is assumed€– a principal axes of the moment tensor are not
value appropriate for volcanic rock at or near exactly [1:1:2], implying additional complex
liquidus temperatures (Murase and McBirney, ity beyond the simple first-order mechanism
1973). A simple crack model, illustrated by the illustrated in Figures 15.8 (a) and 15.8 (b). To
gray-shaded planes in Figures 15.8 (c) and 15.8 investigate this complexity, Chouet et€al. (2008)
(d), therefore constitutes an adequate first-order considered a point source composed of two
representation of the source mechanism produ intersecting cracks and performed a search for
cing the moment components shown in Figures the best-fitting model by systematically varying
15.8 (a) and 15.8 (b). the azimuth ϕ and polar angle θ defining the
(a) (b)
Moment tensor Moment tensor
Figure 15.8╇╇ Source mechanisms
of the two types of explosions
M11 M11
occurring at Stromboli in September
1997. (a) Source-time functions
M22 M22 obtained for a Type-1 event, and
(b) source-time functions derived
M33 M33 for a Type-2 event, in which six
4 � 1012 Nm
4 � 1012 Nm
moment-tensor components and
three single-force components are
M12 M12
assumed for the source mechanism.
(c) Source mechanism imaged for
M23 M23 the Type-1 event, and (d) source
mechanism imaged for the Type-2
M31 M31 event. The reference coordinates
for the eigenvectors are W (west), S
(south), and U (up). The eigenvectors,
Single force Single force marked by bold arrows, have been
F1 F1 normalized to a maximum length of
2 and represent the main deflation
phase of the source seen during the
2 � 108 N
4 � 108 N
F2 F2
interval 7–10 s in the Type-1 event in
(a), or interval 30–40 s in the Type-2
F3 F3 event in (b).
0 10 20 0 20 40 60
Time (s) Time (s)
(c) (d)
U U
φ = 41° S φ = 45° S
W W
1
2 2
1.1
0.8 1
orientation of the normal vector to each crack moment tensor and single force components for
plane (see Eq. (15.8); note that this source inver the two-crack mechanism very closely match
sion solves for a single moment tensor with the moment and single force components illus
shared time history in each crack, plus three trated in Figure 15.8 (a), and the waveform fits
single forces with variable time histories). Figure obtained by these two approaches are virtually
15.9 shows the result of their reconstruction indistinguishable (Chouet et€ al., 2008). A simi
of the source mechanism for the Type-1 event. lar dual crack mechanism was derived for the
This solution points to a dominant crack whose Type-2 event.
orientation is within a few degrees of the crack Although the point sources imaged in the
imaged in Figure 15.8 (a), and subdominant above studies provide a very good match for the
crack with similar dip but distinct azimuth. The VLP waveforms recorded on receivers located in
100 m3
source is a composite of two cracks.
M22 The moment tensor in (a) represents
the sum of the moment tensors
M33 0 10 20 representing each crack, where each
4 � 1012 Nm
Single force 2
F1
S
2
2 � 108 N
F2 2
φ1 = 43° φ2 = 67°
W
F3
0 10 20
Time (s)
the upper part of the edifice, noticeable misfits according to this procedure. The two conduit
were observed to remain on receivers surround structures are obtained simply by extending
ing the edifice near sea level. These misfits were the two crack planes imaged at the upper point
interpreted by Chouet et€al. (2008) as evidence of source up to the surface, extending the plane of
contributions from deeper source components, the dominant crack imaged at the lower point
and to quantify these components they con source up to the upper point source, and viewing
ducted a search for another composite mech the subsidiary crack at the lower point source
anism at a second point source. In this search, as the downward extension of the conduit. The
the positions of the two cracks in the original main branch of conduit activated during erup
source remained fixed, while the second point tions at the northern vent is composed of a dike
source was again assumed to be a composite of dipping steeply to the northwest and extending
two intersecting cracks combined with three essentially straight from 80 m below sea level to
single force components. Once the position and the crater floor, 760 m above sea level (Fig. 15.10
mechanism of the second point source were a). At a depth of 80€m below sea level the conduit
identified, a fine adjustment of the coupled features a sharp corner leading into a dike seg
mechanisms of the two sources was carried out ment dipping to the southeast. The main dike
to find the absolute minimum of residual error segment above, and deep segment below the
between fitted synthetics and data. abrupt corner, both strike northeast-southwest
Figure 15.10 shows the two conduit structures along a direction parallel to the elongation of
compatible with the two point sources (marked the volcanic edifice. A subsidiary dike segment
by white circles) reconstructed from inversion branches off the main crack-like conduit at ele
of waveform data for Type-1 and Type-2 events vations near 440 m (dark green segment in Fig.
a gas slug through the shallow bifurcations in of the same resonator to a sustained excitation.
the two conduits (Fig. 15.10 (c)). In contrast, the LP events are particularly useful in the quantifi
temporal features of the lower point sources cation of magmatic and hydrothermal processes
associated with the conduit corners below sea because the properties of the resonator system
level were interpreted by Chouet et€al. (2008) as at the source of this event can be inferred from
a passive response of the conduit to the action the properties of the decaying oscillations in the
of the upper source. Thus, each discontinuity in tail of the seismogram. The damped oscillations
the conduit provides a site where pressure and in the LP coda are quantified by two parameters,
momentum changes resulting from flow proc T, and Q, where T is the period of the dominant
esses associated with the passage of a gas slug mode of oscillation, and Q is the quality factor
through the discontinuity are coupled to the of the oscillatory system representing the com
Earth, or where the elastic response of the con bined effects of intrinsic and radiation losses.
duit can couple back into pressure and momen Interpretations of the oscillating characteris
tum changes in the fluid. Support for this view tics of LP sources have mostly relied on a model
is provided by laboratory simulations (James of a fluid-driven crack (Chouet, 1986). This
et€ al., 2006) investigating the ascent of a slug model, which has the most natural geometry
of gas in a vertical liquid-filled tube featuring a that satisfies mass-transport conditions at depth
sharp flare (see Fig. 15.6). The pressure pulse in beneath a volcano (pipe-like structures are not
the liquid and downward force on the apparatus expected to exist under the prevailing pressure
holding the tube observed as the slug clears the conditions at depth), is supported by results from
flare are both consistent with the mechanisms inversions of LP waveforms recorded at several
imaged for the upper sources at Stromboli (com volcanoes (Kumagai et€ al., 2002b, 2005; Waite
pare volume changes and vertical force F3 in Fig. et€al., 2008). A simplified two-dimensional model
15.9 with pressure and vertical tube displace of a fluid-driven crack was first introduced by
ment in Fig. 15.6 (b)). The force observed under Aki et€ al. (1977). Although this model included
laboratory conditions also satisfactorily scales both the driving excitation and geometry appro
to the magnitude of the force seen in the vol priate for transport, the fluid inside the crack
canic environment, lending further support for was treated as a passive cushion that did not
this interpretation. support the acoustic propagation of the pres
sure disturbance caused by the motion of the
crack wall. An extension of this model includ
ing active fluid participation was later proposed
15.4╇ ╇Sources of long-period by Chouet and Julian (1985), who considered
seismicity a simultaneous solution of the elastodynam
ics and fluid dynamics for a two-dimensional
The gaseous fluid mixtures (gas, liquid and/or crack. This model was further extended to three
solid particles, and associated mixtures) compos dimensions by Chouet (1986) and was exten
ing volcanic fluids often lead to strong velocity sively studied by Chouet (1988, 1992).
contrasts between the fluid and encasing solid. The three-dimensional model consists of a
This in turn favors the entrapment of acoustic single isolated rectangular crack embedded in
energy at the source, leading to long-lasting an infinite elastic body and assumes zero mass
oscillations commonly observed as LP events and transfer into and out of the crack (Fig. 15.12).
tremor (Chouet, 1996a). The waveform of the LP Crack resonance is excited by a pressure tran
event is characterized by simple decaying har sient applied symmetrically on both walls over
monic oscillations except for a brief interval at a small patch of crack wall. In this model, the
the event onset (Fig. 15.11). This signature may crack aperture is assumed to be much smaller
be viewed as the response of a fluid-filled res than the seismic wavelengths of interest and the
onator to a short-lived excitation. By the same motion of the fluid inside the crack is treated as
token, tremor may be attributed to the response two-dimensional in-plane motion.
d
Fluid: ∂ V l ∂ t = −13ρsρ f ∂ p∂ x l ,
(15.24)
α, ρs ∂ p∂ t = − bµ∂ V m ∂ x m − 2bµ Ldv d ,
(15.25)
a, ρf
where the crack is set in the plane Z = 0 in
a Cartesian coordinate system x,y,z. In these
L Rock equations, nondimensional variables are indi
matrix
cated by an overbar; v̅i and σ̅ij represent the
Fluid
particle velocity and stress components in the
solid, v̅d = v̅z(d/2) is the normal component of
W velocity of the crack wall, V̅l and p̅ are the vel
ocity components and pressure of the fluid.
The parameters ρs and ρf are the densities of
the solid and fluid, μ and b are the rigidity of
Figure 15.12╇╇ Geometry of the fluid-filled crack model of
the solid and bulk modulus of the fluid, and
Chouet (1986). The crack has length L, width W, and aperture
d, and contains a fluid with sound speed a and density ρf. The L and d represent the crack length and crack
crack is embedded in an elastic solid with compressional aperture, respectively. The dimensionless vari
wave velocity α and density ρs. Excitation of the crack is ables are obtained through the following scal
provided by a pressure transient applied symmetrically on ing relations:
both walls over the small areas indicated by the gray patches.
Length x i = x i / L (15.26)
Time t = αt / L
Assuming a Poisson solid (λ = μ), Chouet (15.27)
(1986) calculated the response of the crack by Stress σ ij = σ ij / σ0
solving the following dimensionless equations (15.28)
together with boundary conditions for stresses
Displacement u i = u i µ / ( Lσ 0 )
at the crack surface and fluid flow at the crack (15.29)
perimeter:
Velocity v i = v i µ / (ασ0 ) ,
(15.30)
Solid: ∂ v i ∂ t = 13∂σ ik ∂ x k , (15.22)
where σ0 is the effective stress, and
∂σ ij ∂ t = ∂ v k ∂ x k δ ij + ∂ v i ∂ x j + ∂ v j ∂ x i , (15.23) α = (λ + 2µ) / ρs is the compressional wave vel
ocity in the solid.
These studies demonstrated that the reson in a fluid-filled borehole (Biot, 1952). Unlike
ance of the crack is sustained by a slow wave the tube wave, however, as the wavelength
termed the “crack wave.” This slow wave can be increases to infinity the velocity of the crack
understood from a consideration of Eqs. (15.24) wave approaches zero in inverse proportion to
and (15.25), which can be combined to obtain the square root of wavelength (Ferrazzini and
the relation: Aki, 1987). A brief summary of the basic proper
2 2 2
∂ 2 p∂ t + 2C∂ v d ∂ t = (aα)2 (∂ 2 p∂ x + ∂ 2 p∂ y ) , ties of the crack wave can be found in Kawakatsu
(15.31)
and Yamamoto (2007). Figure 15.13 illustrates
where a = b / ρ f is the acoustic velocity of the the dispersion characteristics of the crack wave
fluid, and C = (b/μ)(L/d) is a dimensionless par in the model of Chouet (1986).
ameter called “crack stiffness’’ (Aki et€al., 1977; Synthetic seismograms calculated with the
Chouet, 1986). Let us consider the simple case fluid-filled crack model bear strong resem
where the normal component of displacement blance to observed LP waveforms (Fig. 15.14),
of the crack wall ud̅ is proportional to p̅, convincingly demonstrating the critical import
ance of active fluid participation in the source
ud = ε p , (15.32)
process of volcanic signals. The slow character
where ε is a nondimensional proportionality istics of the crack wave also lead to more realis
constant. Equation (15.31) may then be recast tic estimates of crack dimensions compared to
as estimates based on the sound speed of a fluid
2 2 2 embedded in a resonator with perfectly rigid
(1 + 2C ε)∂ 2 p∂ t = (aα)2 (∂ 2 p∂ x + ∂ 2 p∂ y ) , (15.33)
walls. Crack lengths estimated from LP data
or equivalently based on this model typically range from tens to
2 2 2 several hundred meters (Saccorotti et€al., 2001;
∂ 2 p∂ t = (aeα)2 (∂ 2 p∂ x + ∂ 2 p∂ y ) . (15.34)
Kumagai et€al., 2002a, 2005).
Equation (15.34) represents the dimensionless As formulated, the fluid-filled crack model
wave equation for pressure with phase velocity accounts for the radiation loss only. Intrinsic
ae = be / ρ f , where be represents an effective losses caused by dissipation mechanisms within
bulk modulus defined as the fluid must be treated separately. Detailed
analyses of the dependence of crack resonance
be = b(1 + 2C ε) . (15.35)
on fluid composition by Kumagai and Chouet
For a perfectly rigid wall (no deformation), (2000) show that the Q factor of the crack res
ε =€0 and the phase velocity ae = a. When the wall onance increases monotonically with increasing
deforms in response to increasing pressure, ε > 0 ratio α/a, and systematic investigations of LP sig
and ae becomes smaller than a through a reduc natures based on their results suggest that dusty
tion of the effective bulk modulus. As shown gases or bubbly basalt are common fluids in LP
in Eq. (15.35), the effective bulk modulus also events of magmatic origin (Gil Cruz and Chouet,
decreases with increasing crack stiffness, dem 1997; Kumagai and Chouet, 1999), and that wet
onstrating the critical importance of this param gases, steam, and bubbly water are typically rep
eter in controlling the crack dynamic behavior. resentative of the source of LP events of hydro
The asymptotic behavior of the crack wave thermal origin (Saccorotti et€al., 2001; Kumagai
was investigated analytically by Ferrazzini and et€al., 2002a).
Aki (1987) in a study of normal modes trapped The model of Chouet (1986) does not address
in a liquid layer sandwiched between two elas the excitation mechanism of LP events or
tic half spaces. Ferrazzini and Aki (1987) showed tremor. Rather, the spatio-temporal properties
that crack wave speed increases with decreasing of the pressure transient triggering the crack
wavelength and in the short-wavelength limit resonance are preset as kinematic conditions
reduces to the Stoneley wave propagating along in the model. The usefulness of this model is
the fluid–solid interface. These properties are thus restricted to a quantification of the crack
analogous to those of tube waves propagating resonance and properties of the fluids at the
0.6
b/µ = 0.1
α /a = 4
0.5
0.4
Velocity ratio, v/a
0.3
0.2
C = 50
C = 100
0.1
C = 200
C = 500
0
0 0.2 0.4 0.6 0.8 1.0
Dimensionless wavelength, Λ/L
LP source is located very close to the magma address the frequency-dependent properties of
conduit, LP activity is also seen to trigger a pas materials (i.e., bubbly liquids) found in volcanic
sive response of the conduit itself (Waite et€al., and hydrothermal processes.
2008). In contrast, LP seismicity observed at Neuberg et€al. (2000) and Jousset et€al. (2003)
Redoubt was interpreted as a more energetic investigated seismic-acoustic wave conver
form of magmatic-hydrothermal interaction sion and coupling in a shallow rectangular
where an unsteady choked flow of magmatic conduit embedded in a homogeneous elastic
gases provides a natural source of pressure per half space. They used a 2D finite-difference
turbation at the origin of LP events and tremor scheme to model major features of the LP
(Chouet et€ al., 1994; Morrissey and Chouet, seismic wave field and study the behavior of
1997). Magmatic LP events produced during single LP events as well as tremor. In their
bursts of ash-laden gases associated with vulca model, seismic propagation in the elastic solid
nian activity at Galeras (Gil Cruz and Chouet, and acoustic propagation in the fluid-filled
1997), and Popocatépetl (Arciniega-Ceballos conduit are solved simultaneously using a
et€ al., 2008) involve a pumping mechanism in single velocity-stress computational scheme.
shallow fractures similar to that inferred for The fluid is defined by a zero shear-velocity
hydrothermal LP events at Kusatsu-Shirane, (rigidity μ = 0), and appropriate values for the
Kilauea, and Mount St. Helens. LP events density and sound speed (compressional wave
accompanying endogeneous dome growth at velocity, α). This approach does not require
Soufrière Hills Volcano, Montserrat, appear to explicit boundary conditions at the conduit
be related to a more complex process involving wall to define the coupling between the fluid
the production of shear fractures in highly vis and solid. Acoustic-seismic conversion results
cous magma at the glass transition, injection of from energy transmission controlled by effect
dusty gases into these fractures, and resonance ive material properties at the fluid–solid inter
of the fracture network and/or possibly reson face. Using a 2D finite-difference scheme and
ance of a bubble-rich magma excited by the rectangular conduit model, Jousset et€al. (2004)
energy release from the brittle failure (Neuberg further investigated the effects of viscoelas
et€ al., 2006). LP events in the basaltic systems ticity and topography on the amplitudes and
at Kilauea (Ohminato et€al., 1998; Chouet, 2003) spectra of LP events. Their study indicates
and Stromboli (Chouet et€ al., 2003, 2008) are that the effects of anelastic attenuation and
attributed to pressure disturbances generated topography can induce significant distortion
during the transit of large slugs of gas through in LP spectra. Their results also suggest that
conduit discontinuities. the rheological properties of magmas may
In the model of Chouet (1986), a simultan be constrained from detailed analyses of LP
eous solution of the equations of motion of seismograms.
the fluid inside the crack and the surrounding A nonlinear excitation mechanism of LP seis
elastic solid (Eqs. (15.22)€ –(15.25)) is obtained micity by fluid flow was proposed by Julian (1994).
numerically by using a 3-D time-domain Using a simple lumped-parameter model, Julian
finite-difference scheme. Another promising investigated the elastic coupling of the fluid and
approach to numerically simulate the dynamic solid as a means to produce self-excited oscilla
response of a fluid-filled crack is to formu tions in a viscous incompressible liquid flowing
late this problem using the frequency-domain through a channel with compliant walls. In his
boundary integral method (Yamamoto and model, an increase in flow velocity leads to a
Kawakatsu, 2008). The boundary integral decrease of fluid pressure via the Bernoulli effect.
method relies on a formulation of governing As a result, the channel walls move inward and
equations in terms of integrals on boundaries, constrict the flow, causing an increase in fluid
which can be solved by using a point collocation pressure and forcing the channel open again.
method (Yamamoto and Kawakatsu, 2008). This The cyclic repetition of this process is the source
approach appears particularly well-suited to of sustained oscillations in Julian’s model. Julian
Balmforth, N. J., R. V. Craster and A. C. Rust (2005). Chouet, B., P. Dawson and M. Martini (2008).
Instability in flow through elastic conduits and Shallow-conduit dynamics at Stromboli Volcano,
volcanic tremor, Journal of Fluid Mechanics, 527, Italy, imaged from waveform inversion. In Fluid
353–377. Motions in Volcanic Conduits: A Source of Seismic and
Ben-Menahem, A. and S. J. Singh (1981). Seismic Waves Acoustic Signals, ed S. J. Lane and J. S. Gilbert.
and Sources. New York: Springer-Verlag, 1108 pp. Geological Society, London, Special Publications,
Biot, M. A. (1952). Propagation of elastic waves in a 307, 57–84.
cylindrical bore containing a fluid, Journal of Applied Chouet, B. A., P. B. Dawson, M. R. James and S.
Physics, 23, 997–1005. J. Lane (2010). Seismic source mechanism of
Chouet, B. (1986). Dynamics of a fluid-driven crack degassing bursts at Kilauea Volcano, Hawaii:
in three dimensions by the finite difference Results from waveform inversion in the 10–50 s
method, Journal of Geophysical Research, 91, band, Journal of Geophysical Research, 115, B09311,
13,967–13,992. doi:10.1029/2009JB006661.
Chouet, B. (1988). Resonance of a fluid-driven crack: Dawson, P. B., B. A. Chouet and J. Power (2011).
Radiation properties and implications for the Determining the seismic source mechanism
source of long-period events and harmonic tremor, and location for an explosive eruption with
Journal of Geophysical research, 93, 4375–4400. limited observational data: Augustine Volcano,
Chouet, B. (1992). A seismic source model for the Alaska, Geophysical Research Letters, 38, L03302,
source of long-period events and harmonic tremor. doi:10.1029/2010GL045977.
In IAVCEI Proceedings in Volcanology, Volume 3, ed Ekström, G., M. Nettles and G. A. Abers (2003).
P. Gasparini, R. Scarpa and K. Aki. New York: Glacial earthquakes, Science, 302, 622–624,
Springer-Verlag, pp. 133–156. doi:10.1126/science.1088057.
Chouet, B. (1996a). Long-period volcano seismicity: Eshelby, J. D. (1957). The determination of the
its source and use in eruption forecasting, Nature, elastic field of an ellipsoidal inclusion and related
380, 309–316. problems. Proceedings of the Royal Society, A241,
Chouet, B. (1996b). New methods and future trends 376–396.
in seismological volcano monitoring. In Monitoring Fehler, M. (1983). Observations of volcanic tremor
and Mitigation of Volcano Hazards, ed. R. Scarpa and at Mount St. Helens volcano, Journal of Geophysical
R. I. Tilling. New York: Springer-Verlag, pp. 23–97. Research, 88, 3476–3484.
Chouet, B. (2003). Volcano seismology, Pure and Ferrazzini, V. and K. Aki (1987). Slow waves trapped
Applied Geophysics, 160, No. 3–4, 739–788. in a fluid-filled infinite crack: Implications for
Chouet, B. and B. R. Julian (1985). Dynamics of an volcanic tremor, Journal of Geophysical Research, 92,
expanding fluid-filled crack, Journal of Geophysical 9,215–9,233.
Research, 90, 11,187–11,198. Gil Cruz, F., and B. A. Chouet (1997). Long-period
Chouet, B. A., R. A. Page, C. D. Stephens, J. C. Lahr events, the most characteristic seismicity
and J. A. Power (1994). Precursory swarms of accompanying the emplacement and extrusion of
long-period events at Redoubt Volcano (1989– a lava dome in Galeras Volcano, Colombia, in 1991,
1990), Alaska: Their origin and use as a forecasting Journal of Volcanolology and Geothermal Research, 77,
tool, Journal of Volcanology and Geothermal Research, 121–158.
62, 95–135. James, M. R., S. J. Lane, B. Chouet and J. S. Gilbert
Chouet, B., P. Dawson T. Ohminato et€al. (2004). Pressure changes associated with the ascent
(2003). Source mechanisms of explosions at and bursting of gas slugs in liquid-filled vertical
Stromboli Volcano, Italy, determined from and inclined conduits, Journal of Volcanolology and
moment-tensor inversions of very-long-period Geothermal Research, 129, 61–82.
data, Journal of Geophysical Research, 108(B1), 2019, James, M. R., S. J. Lane, and B. A. Chouet (2006). Gas
doi:10.1029/2002JB001919. slug ascent through changes in conduit diameter:
Chouet, B., P. Dawson and A. Arciniega-Ceballos Laboratory insights into a volcano-seismic
(2005). Source mechanism of Vulcanian degassing source process in low-viscosity magmas,
at Popocatepetl Volcano, Mexico, determined Journal of Geophysical Research, 111, B05201,
from waveform inversions of very long period doi:10.1029/2005JB003718.
signals, Journal of Geophysical Research, 110, B07301, James, M. R., S. J. Lane and S. B. Corder (2008).
doi:10.1029/2004JB003524. Modelling the rapid near-surface expansion of
gas slugs in low-viscosity magmas. In Fluid Motions Kumagai, H. (2006). Temporal evolution of a
in Volcanic Conduits: A Source of Seismic and Acoustic magmatic dike system inferred from the complex
Signals, ed S. J. Lane and J. S. Gilbert. Geological frequencies of very long period seismic signals,
Society, London, Special Publications, 307, Journal of Geophysical Research, 111, B06201,
147–167. doi:10.1029/2005JB003881.
Johnson, J. B., J. M. Lees, A. Gerst, D. Sahagian and Kumagai, H. and B. A. Chouet (1999). The complex
N. Varley (2008). Long-period earthquakes and frequencies of long-period seismic events as probes
co-eruptive dome inflation seen with particle of fluid composition beneath volcanoes, Geophysical
image velocimetry, Nature, 456, 377–381. Journal International, 138, F7€– F12.
Jousset, P., J. Neuberg and S. Sturton (2003). Kumagai, H. and B. A. Chouet (2000). Acoustic
Modelling the time-dependent frequency content properties of a crack containing magmatic or
of low-frequency volcanic earthquakes, Journal of hydrothermal fluids, Journal of Geophysical Research,
Volcanology and Geothermal Research, 128, 201–223. 105, 25 493–25 512.
Jousset, P., J. Neuberg and A. Jolly (2004). Modelling Kumagai, H., B. A. Chouet and M. Nakano (2002a).
the low-frequency volcanic earthquakes in a Temporal evolution of a hydrothermal system in
viscoelastic medium with topography, Geophysical Kusatsu-Shirane Volcano, Japan, inferred from
Journal International, 159, 776–802. the complex frequencies of long-period events,
Julian, B. R. (1994). Volcanic tremor, nonlinear Journal of Geophysical Research, 107(B10), 2236,
excitation by fluid flow, Journal of Geophysical doi:10.1029/2001JB000653.
Research, 99, 11,859–11,877. Kumagai, H., B. A. Chouet and M. Nakano (2002b).
Kanamori, H. and J. W. Given (1982). Analysis of Waveform inversion of oscillatory signatures
long-period waves excited by the May 18, 1980, in long-period events beneath volcanoes,
eruption of Mount St. Helens€– A terrestrial Journal of Geophysical Research, 107(B11), 2301,
monopole? Journal of Geophysical Research, 87, doi:10.1029/2001JB001704.
5,422–5,432. Kumagai, H., K. Miyakawa, H. Negishi et€al. (2003).
Kanamori. H., and J. Mori (1992). Harmonic Magmatic dike resonances inferred from
excitation of Mantle Rayleigh waves by the very-long-period seismic signals, Science, 299,
1991 eruption of Mount Pinatubo, Philippines, 2058–2061.
Geophysical Research Letters, 19, 1,856–1,866. Kumagai, H., B. A. Chouet and P. B. Dawson (2005).
Kanamori, H., J. W. Given and T. Lay (1984). Analysis Source process of a long-period event at Kilauea
of seismic body waves excited by the Mount Volcano, Hawaii, Geophysical Journal International,
St. Helens eruption of May 18, 1980, Journal of 161, 243–254.
Geophysical Research, 89, 1856–1866. Lane, S. J., B. A. Chouet, J. C. Phillips et€al. (2001).
Kawakatsu, H. (1989). Centroid single force inversion Experimental observations of pressure oscillations
of seismic waves generated by landslides, Journal of and flow regimes in an analogue volcanic system,
Geophysical Research, 94, 12 363–12 374. Journal of Geophysical Research, 106, 6461–6476.
Kawakatsu, H. and M. Yamamoto (2007). Volcano Latter, J. H. (1979). Volcanological observations
seismology. In Treatise on Geophysics Volume 4, at Tongariro National Park, 2, Types and
Earthquake Seismology, ed. H. Kanamori and G. classification of volcanic earthquakes, 1976–
Schubert. New York: Elsevier, pp. 389–420. 1978, New Zealand Department of Science and
Kawakatsu, H., S. Kaneshima, H. Matsubayashi Industrial Research, Geophysical Division, Wellington,
et€al. (2000). Aso94: Aso seismic observation with Report 150, 60 pp.
broadband instruments, Journal of Volcanology and Linde, A. T., K. Agustsson, I. Selwyn Sacks and
Geothermal Research, 101, 129–154. R. Stefansson (1993). Mechanism of the 1991
Knopoff, L. and M. J. Randall (1970). The eruption of Hekla from continuous borehole strain
compensated linear-vector dipole: A possible monitoring, Nature, 365, 737–740.
mechanism for deep earthquakes, Journal of McNutt, S. R. (2002). Volcano seismology and
Geophysical Research, 75, 4957–4963. monitoring for eruptions. In International Handbook
Konstantinou, K. I. and V. Schlindwein (2002). of Earthquake and Engineering Seismology, Part A,
Nature, wavefield properties and source Chapter 25, ed. W. H. K. Lee, H. Kanamori, P. C.
mechanism of volcanic tremor: a review, Journal of Jennings and C. Kisslinger. New York: Academic
Volcanology and Geothermal Research, 119, 161–187. Press, pp. 383–406.
McNutt, S. R. (2005). Volcanic seismology, Annual components observed during the Asama 2004
Review of Earth and Planetary Sciences, 32, 461–491. volcanic activity in Japan, Earth, Planets and Space,
Mogi, K. (1958). Relation between the eruptions 58, 583–593.
of various volcanoes and the deformations of Rust, A. C., N. J. Balmforth and S. Mandre (2008). The
the ground surfaces around them, Bulletin of the feasibility of generating low-frequency volcano
Earthquake Research Institute, 36, 99–134. seismicity by flow through a deformable channel.
Morrissey, M. M. and B. A. Chouet (1997). A In Fluid Motions in Volcanic Conduits: A Source of Seismic
numerical investigation of choked flow dynamics and Acoustic Signals, ed S. J. Lane and J. S. Gilbert.
and its application to the triggering mechanism Geological Society, London, Special Publications,
of long-period events at Redoubt Volcano, Alaska, 307, 45–56.
Journal of Geophysical Research, 102, 7,965–7,983. Saccorotti, G., B. Chouet and P. Dawson (2001).
Murase, T. and A. R. McBirney (1973). Properties Wavefield properties of a shallow long-period
of some common igneous rocks and their melts event and tremor at Kilauea Volcano, Hawaii,
at high temperatures, Geological Society of America Journal of Volcanology and Geothermal Research, 109,
Bulletin, 84, 3563–3592. 163–189.
Nakano, M., H. Kumagai and B. A. Chouet (2003). Sneddon, I. N. and M. Lowengrub (1969). Crack
Source mechanism of long-period events at Problems in the Classical Theory of Elasticity, New York:
Kusatsu-Shirane Volcano, Japan, inferred from John Wiley.
waveform inversion of the effective excitation Takei, Y. and M. Kumazawa (1994). Why have the
functions, Journal of Volcanology and Geothermal single force and torque been excluded from
Research, 122, 149–164. seismic source models?, Geophysical Journal
Neuberg, J., R. Luckett, B. Baptie and K. Olsen (2000). International, 109, 163–189.
Models of tremor and low-frequency earthquake Uhira, K. and M. Takeo (1994). The source of
swarms on Montserrat, Journal of Volcanology and explosive eruptions of Sakurajima Volcano, Japan,
Geothermal Research, 101, 83–104. Journal of Geophysical Research, 99, 17 775–17 789.
Neuberg, J. W., H. Tuffen, L. Collier et€al. (2006). The Waite, G. P., B. A. Chouet and P. B. Dawson (2008).
trigger mechanism of low-frequency earthquakes Eruption dynamics at Mount St. Helens imaged
on Montserrat, Journal of Volcanology and Geothermal from broadband seismic waveforms: Interaction
Research, 153, 37–50. of the shallow magmatic and hydrothermal
Nishimura, T. (1995). Source parameters of volcanic systems, Journal of Geophysical Research, 113, B02305,
eruption earthquakes at Mount Tokachi, doi:10.1029/2007JB0052592007.
Hokkaido, Japan, and magma ascending model, Yamamoto, M., and H. Kawakatsu (2008). An efficient
Journal of Geophysical Research, 100, 12465–12473. method to compute the dynamic response of a
Nishimura, T. and H. Hamaguchi (1993). Scaling fluid-filled crack, Geophysical Journal International,
law of volcanic explosion earthquake, Geophysical 174, 1174–1186.
Research Letters, 20, 2479–2482.
Nishimura, T., H. Nakamichi, S. Tanaka et€al. (2000).
Source process of very long period seismic events
associated with the 1998 activity of Iwate Volcano,
northeastern Japan, Journal of Geophysical Research, Exercises
105, 19 135–19 147.
Ohminato, T. and B. A. Chouet (1997). A free-surface 15.1 Use the expression for the moment density
boundary condition for including 3D topography tensor m pq = [ u i ]ν j [ λδ ij δ pq + µ (δ ip δ jq + δ iq δ jp )] in Eq.
in the finite-difference method, Bulletin of the (15.2) to demonstrate that a point source repre
Seismological Society of America, 87, 494–515. senting a thin crack opening in the direction ν
Ohminato, T., B. A. Chouet, P. B. Dawson and S. Kedar given by the angles θ and ϕ (see Fig. 15.3 (a)) has
(1998). Waveform inversion of very-long-period the moment tensor given in Eq. (15.8). Show
impulsive signals associated with magmatic that the eigenvalues of the tensile crack con
injection beneath Kilauea volcano, Hawaii, Journal sist of three dipoles with magnitudes λΔV, λΔV,
of Geophysical Research, 103, 23 839–23 862. and (λ + 2μ)ΔV, where the dominant dipole is
Ohminato, T., M. Takeo, H. Kumagai et€al. (2006). oriented normal to the crack plane and ΔV is
Vulcanian eruptions with dominant single force the volume change of the crack.
15.2 Use the formula for Mpq in Eq. (15.2) to obtain couple) as illustrated in Fig.€15.2 (b), where each
the moment tensor in Eq. (15.9) for a point couple shares the same magnitude µ[ u1 ] S ,
source representing the radial expansion of in which S is the fault area and [ u1 ] is final
a cylinder with axis orientation as given in value of slip averaged over the fault surface.
Fig. 15.4 (a). Show that the eigenvalues of a Use an eigenvalue decomposition to show that
pipe consist of three dipoles with magnitudes this double couple is equivalent to two vector
λΔV, (λ + μ)ΔV, and (λ + μ)ΔV, where the small dipoles equal in magnitude and opposite in
est dipole is oriented along the pipe axis sign and that this source has no volumetric
(Fig. 15.4 (b)) and ΔV is the volume change of component.
the pipe.
15.3 Assume a point source of shear on a horizontal
Online resources available at www.cambridge.
fault plane ε3 = 0 with slip [u1] parallel to the
org/fagents
fault as shown in Fig. 15.2(a). Use Eq. (15.2) to
show that the equivalent force system for this • Glossary
source consists of two couples (i.e., a double • Additional reading
Volcano acoustics
Milton A. Garcés, David Fee, and Robin Matoza
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K.╯P. Gregg, and Rosaly M.╯C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
c = λf, we deduce that infrasonic frequencies cor- situations, the wave equation (Eq. (16.1)) does
respond to spatial scales on the order of tens of not hold and it is necessary to revert back to the
meters or larger. Audible sounds from volcanoes more fundamental hydrodynamic equations.
represent the veritable tip of the iceberg com- However, in practice, a distance from the source
pared to the majority of acoustic energy radiated can be defined where P drops below equilibrium
at infrasonic frequencies. For example, sounds values and the acoustic equations are valid.
from strombolian explosions recorded close to Beyond this distance from the source it is pos-
the vent have frequencies between 1 Hz and 10 sible to construct equivalent source models that
kHz (Garcés, 1995), with a range of eight orders capture the main features of an acoustic source
of magnitude in acoustic intensity. Because process. In this chapter we concentrate on the
infrasound is efficiently generated by large-scale simplest and most general relationships that
eruption processes and may travel through the may permit reasonable physical interpretations
atmosphere for large distances (>€1000 km) with of field data and comparisons between different
minimal attenuation, the infrasound frequency eruption processes.
band is particularly well suited to the remote
monitoring of eruption signals. 16.2.1â•… Sound pressure levels
The field of acoustics is a branch of fluid One of the most fundamental acoustic measure-
mechanics. The equations of motion for acoustic ments is pressure amplitude (Eq. (16.2)). To date,
propagation are a special case of the linearized recorded volcano acoustic signals span over 14
Navier–Stokes equations for irrotational, com- orders of magnitude in sound intensity, so it is
pressible, Newtonian fluids, subject to a suite of useful to adopt a logarithmic scale known as the
simplifying assumptions. For further details, see sound pressure level (SPL):
Morse and Ingard (1968), Lighthill (1978), Pierce
(1981), and Kinsler et€al. (1982). p2 p
A propagating sound wave obeys the wave SPL = 10 log 10 rms
2
= 20 log 10 rms , (16.3)
pref p ref
equation given by (Morse and Ingard, 1968):
where
1 ∂2 p
∇2 p − = 0, (16.1)
c 2 ∂t 2 1 Ts
= p2 = ∫
2
p rms p 2 (t)dt, (16.4)
Ts 0
where c is the sound speed, and the acoustic
pressure p is defined as the difference between and p is the measured acoustic pressure over a
the instantaneous pressure P and the ambient specified frequency band, pref is a reference pres-
(or equilibrium) pressure P0, sure, and Ts is a chosen time interval for inte-
p = P − P0 . gration. Root-mean-square (rms) pressure (prms)
(16.2) estimates are robust when the acoustic signal
Thus, the governing equations in acoustics are is stationary, meaning that its statistical proper-
perturbation solutions, and exclude steady state ties do not change substantially over the period
or (incompressible) Bernoulli flow where P€– P0 of integration. For a continuous signal with a
would be time-invariant. Only unsteady processes constant period, prms can be approximated by
produce sound. Of particular relevance to vol- prms2 = pmax2/2, where pmax is the peak pressure
canological applications is the fundamental in Pa. However, for impulsive, transient signals
assumption that acoustic pressure and density such as explosions, the integration interval
fluctuations from equilibrium values are small. should match the signal duration to provide a
Earth’s atmospheric equilibrium pressure (P0) sound exposure level Ld = SPL + 10 log10Ts. The
is ~0.1 MPa at sea level. In close proximity to reference level of sound exposure in air is (20
powerful volcanic sources, acoustic pressure μPa)2 s (ANSI S12.7–1986,1986). The SPL or Ld are
contributions to P can approach or even exceed not meaningful without a specification of the
the ambient value of P0 (Eq. (16.2)). In these frequency band used to compute Eq. (16.4), and
a correction for the data acquisition response in where ps is the acoustic source pressure and Ω
that band (ANSI S1.13–2005, 2005). is the solid angle of a surface. This power esti-
Sound intensity, I, in W m−2, is a measure of mate is time-invariant for sustained, statistically
the average rate of acoustic energy flow through stationary excitations, so it is particularly use-
an area normal to the propagation direction, ful when comparing the energetics of processes
and is proportional to the square of the pres- that are stationary over the window of integra-
sure far from the source. The intensity can be tion in Eq. (16.4). However, for transient events
estimated from: where the interval of integration Ts is restricted
to the time of the main pulse, a more useful
1 Ts p2 estimate is the total energy (E) radiated by the
I≅
ρcTs ∫ 0
p 2 (t)dt =
ρc
, (16.5)
event, in Joules. The energy of a transient event
within the event time window can be estimated
where the sound speed c and the equilibrium by E = Ts Π.
density ρ of Earth’s atmosphere at 20 °C and It is assumed that the acoustic radiation
sea level pressures are 343 m s−1 and 1.2 kg m−3, features of the source do not change over the
respectively. Because there is only a propor- time integration implicit in the rms estimate.
tionality factor (ρc) between intensity and the Because source directivity is only a function of
square of pressure, SPL estimates can be read- angle, radial and angular variables may be sepa-
ily converted to intensity by using a reference rated as:
intensity level of 10–12 W m−2 in air. Intensity
estimates are useful for comparing the acoustic p s (r , Ω) = pax (r ) H (Ω), (16.7)
response of different environments. Far from a
source, the ratio of the acoustic pressure p to the where pax(r) contains the range dependence
associated fluid (particle) velocity u is equal to and H contains all the angular directivity infor-
ρc, and is referred to as the characteristic acous- mation. The simplest relationship for radial
tic impedance. We note that the proportionality dependence is obtained when we assume that
between acoustic pressure and particle velocity intensity decreases only as the inverse square
(p = ρcu) shows that compressibility induces pro- of radial distance, a relationship known as the
portionally higher pressure changes than steady spherical spreading law:
incompressible subsonic flow velocity U < c,
where (from Bernoulli’s equation) P ~ ρU2. r 2 ps2 (r ) = const. (16.8)
16.2.2â•… Source energy and power In this general case, the source level (SL) at a
In source modeling research, it is often useful standard r = 1 m from the source can be esti-
to estimate how powerful a sound is at the vol- mated from:
canic source. Source levels (SL) may be estimated
by removing propagation and source directiv- pax
2
(r = 1) p
SL ≡ 10 log 10 2
= 20 log 10 rms (16.9)
ity effects, a task that in practice is fraught pref p ref
with complications. Although we would like to + 20 log 10 R = SPL + 20 log 10 R,
remove atmospheric and boundary (e.g., top-
ography) effects, these effects are notoriously where R is the range in meters to the source
unstable near volcanoes and lead to additional and prms is the pressure at the receiver. Note that
variability in source level estimates. there is a predicted 6 dB drop per doubling of
Total radiated power, Π (in Watts), may be range to the source. Likewise, an increase by a
estimated by integrating the intensity (Eq. (16.5)) factor of two in pressure is equivalent to a SPL
over a surface enclosing the source: increase of 6 dB. The quantity 20log10R is the sim-
plest form of transmission loss (TL), and provides
1
ρc ∫ Ω
∏s = p s2 (r , Ω) r 2dΩ, (16.6) a measure of intensity decrease due to spherical
spreading. If sound is trapped in atmospheric
channels defined by Earth’s temperature and acoustic pressure per frequency band in a time
wind stratification (Brown and Garcés, 2008), window. The decibel unit for PSD is dB rela-
cylindrical spreading would produce a geomet- tive to (20 μPa)2 Hz−1, which is referred to as
rical transmission loss of 10log10R. One may the spectral level. A spectrogram is essentially a
expect observed transmission losses to be some- time-stepping PSD, and is useful for interpret-
where in between the idealized values of cylin- ing time-varying signals. To convert spectral
drical or spherical spreading, as observations are levels to sound pressure levels, a particular
typically affected by topography, vertical tem- frequency band has to be defined. If a spec-
perature and wind gradients, absorption, and tral level is a constant S (dB) over a frequency
scattering from atmospheric turbulence, among band ∆f, then sound pressure level can be esti-
other factors (ANSI S2.20–1983, 1983; Bowman mated from SPL = S + 10 log (∆f). If the PSD is
et€ al., 2004). Nevertheless, Eq. (16.9) permits not constant, then it is necessary to either use
estimates of relative source levels referred to a the mean value over the interval or integrate
calibrated microphone and range. For observa- over the frequency band of interest to get an
tions at a single volcano, Eq. (16.9) can be used estimate of the mean square pressure.
to compare a variety of recorded signals with a PSDs and spectrograms are routinely used
well-characterized reference event. in acoustics for visualizing the dynamic range,
quality, and character of a signal. They com-
16.2.3â•… The spectral domain monly reveal spectral features hidden in wave-
The period of an acoustic wave is the time inter- form displays. Figure 16.1 shows PSDs for a
val between adjacent pressure maxima in a time wide range of sustained eruptive signals; Figure
record, or the duration of a single wavelet cycle 16.1(b) shows low-level activity in Kilauea and
(wavelength). For a simple harmonic wave with Tungurahua volcanoes, with well-defined spec-
a single frequency f, the acoustic pressure as a tral peaks. Figure 16.1(a) shows the broad spec-
function of time would be: trum of more energetic eruptions, with energy
well above background levels (solid line). At a sin-
p(t) = A sin(2π ft − φ ), (16.10)
gle glance it is possible to distinguish between
where f = 1/T, T is the period, t is time in sec- mild and energetic eruptions, suggesting dif-
onds, ϕ is the phase, and A is the amplitude of ferent physical processes are at play. Table 16.1
the sound wave. presents estimates of sustained eruption source
The period of a real signal may be difficult to levels derived from Figure 16.1. These estimates
determine precisely because signals commonly vary by 120 dB (106 Pa), yet only represent the
consist of a superposition of pulses of varying contributions of oscillations with pressure fluc-
periods arriving at different times. A signal tuations in the 1–2 Hz octave band and are not
p(t) consisting of a sum of arbitrary individual a measure of the total source power.
period or frequency (harmonic) components can Because infrasounds are the dominant form
be divided into its individual harmonic compo- of long-range acoustic radiation from volcanoes,
nents through Fourier decomposition. The dis- the rest of this chapter discusses the techniques
crete Fourier transform Fp of the sound pressure used to record and interpret infrasounds and
record p(t) is given by presents a variety of volcanic processes that
may be used to model them.
∞
Fp ( f ) = ∫ −∞
p(t)e − j (2 π f ) t dt,
(16.11)
which has units of Pa Hz−1. The power spectral 16.3╇ ╇ Capture
density (PSD) of a digital signal (Oppenheim
and Schafer, 1989), with units of Pa2 Hz−1, can The goal of many volcano acoustic field cam-
be derived from the discrete version of the paigns is to capture volcanic sound with min-
Fourier transform over a finite time window imal distortion in order to accurately model the
and provides an estimate of the mean squared source. This requires a reliable data acquisition
Table 16.1╇ ╇ Estimated range, sound pressure levels (SPL), spherical spreading transmission loss (TL), and source
level (SL; Eq. (16.9)) between 1 and 2 Hz for sustained volcanic eruptions, interpolated from Figure 16.1.
Range SPL TL SL SL
Source Activity (km) (dB) (dB) (dB) (Pa rms) Comments
Median None N/A 45 N/A 45 4 × 10 –3 Ambient noise model.
Pu’u ‘O’o, magma 2.4 57 68 125 36 4/21/07. Underestimate,
Kilauea, degassing majority of acoustic energy
USA below 1 Hz.
Halemaumau, magma 7 50 77 127 45 3/19/08. Underestimate,
Kilauea, degassing majority of acoustic
USA energy below 1 Hz. Misses
harmonic structure.
Tungurahua, strombolian 37 45 91 136 1.3 × 102 7/29/06. Background
Ecuador open-vent activity.
Mount St. phreatic 13.4 65 83 148 5 × 102 3/9/05. Recorded for over
Helens, 400 km.
USA
Tungurahua, stombolian/ 37 65 91 156 1.3 × 103 5/12/06. Misses harmonic
Ecuador vulcanian structure and
underestimates explosions.
Tungurahua, plinian 37 75 91 166 4 × 103 8/17/06. Underestimate,
Ecuador majority of acoustic energy
below 1 Hz.
edifice are exposed to strong and variable tur- source. Deployments several kilometers from
bulent pressures induced by winds, which are exploding sources also place resources at safer
recorded on the sensors as wind noise that can distances from flying debris and facilitate main-
obscure signals of interest. Although spatial tenance, reducing data loss at critical moments.
wind filters can provide some defense (Hedlin At distances from a few kilometers to thousands
and Raspet, 2003), the most important factor in of kilometers, or in situations where numerous
wind-noise reduction that can realistically be acoustic sources are present, more sophisticated
controlled is the site selection. Locating sensors systems are advisable. Generally, the best design for
in a wind-protected site such as a forest or the remote infrasound monitoring consists of a num-
lee of a topographical barrier is usually worth ber of calibrated sensors precisely time-stamped
the gain in wind-noise reduction in exchange and arranged spatially as arrays. A single infra-
for the disadvantage of being further from the sonic array provides speed and direction of signal
arrivals and can discriminate between dissimilar pressure levels (SPL), spherical spreading trans-
competing sources, whereas two properly sited mission loss (TL), and source level (SL) for some
arrays can locate and identify an arbitrary source of the sustained eruption signatures discussed
(Olson and Szuberla, 2008; Cansi and Le Pichon, herein.
2008). Infrasound arrays are used routinely by the
international infrasound community for detection, 16.4.1â•… Volcano acoustic nomenclature
location, identification, and monitoring of natural In volcano seismology (Chapter 15), tremor
and man-made events (e.g., Garcés et€al., 2004). is a catch-all term used to describe continu-
ous vibration of the ground lasting from
minutes to years. It is possible that different
tremor-generating mechanisms exist at differ-
16.4╇ ╇The volcanic symphony ent volcanoes and even at the same volcano,
as represented by the broad range in temporal
The captivating diversity of outgassing and fluid
and spectral characteristics of observed tremor.
dynamic processes occurring at volcanoes pro-
Similarly, continuous vibration of the air by
duces a rich variety of acoustic signals. Here we
volcanoes is referred to as infrasonic tremor.
present an overview of acoustic observations of
Descriptive terms such as harmonic (with mul-
volcanic eruptions. For a more comprehensive
tiple spectral peaks), spasmodic (with amplitude
listing of volcano acoustic studies, the reader is
variations), episodic (cyclical), and broadband
referred to Harris and Ripepe (2007) and Johnson
(covering a wide frequency range) are used to
and Ripepe (2011).
further describe infrasonic tremor. Harmonic
A general linear model for a recorded vol-
tremor is particularly intriguing as its spec-
cano acoustic signal is given by:
tral signature is reminiscent of musical instru-
p(t) = s(t) * l (t) * g (t), ments, and thus suggestive of ordered spatial
(16.12)
structures within volcanic systems. As for seis-
where p(t) is the observed acoustic signal, s(t) mic volcanic tremor, several candidate infra-
is the source-time function or actual pressure sonic tremor source processes can be invoked;
time-history at the volcano (excitation mech- these are discussed in Sections 16.5 and 16.7.
anism), l(t) denotes local resonance effects Coupling to the atmosphere is necessary for
(e.g., resonance in fluid-filled cavities, cracks, infrasonic tremor, or any other infrasonic
conduits, etc.; Garcés, 1997), g(t) describes all signature.
propagation from the source to the recording Volcano seismic signals have also been clas-
site, and * denotes convolution (Oppenheim and sified based on their period (Chapter 15): very
Schafer, 1989). In theory, g(t) includes all propa- long period (VLP) events have a period of ~2–100
gation effects in the atmosphere. For shallow s (0.01–0.5 Hz); long period (LP) with a period
buried acoustic sources, g(t) can also include ~0.2–2 s (0.5–5 Hz); and short period (SP) below
seismoacoustic coupling through near-surface ~0.2 s (> 5 Hz). LPs and VLPs are commonly
permeable material (e.g., Matoza et€al., 2009a). In attributed to volumetric sources associated
Sections 16.5–16.7 we take a simplified view of with fluid-filled conduits and cracks, whereas
each of these contributions. We deal with source SP quakes are typically attributed to brittle frac-
pressure excitation mechanisms in Section 16.5, ture mechanisms (volcano-tectonic earthquakes,
discuss acoustic propagation away from the Chapter 15). Explosions (Section 16.5) may vary
source in Section 16.6, and in Section 16.7 we substantially in their intensity and duration,
provide an overview of resonance effects in vol- and thus may have energy contributions in any
canic fluid systems. In this section, we provide of these frequency bands. Although volcano seis-
an introduction to the range of acoustic signal mology nomenclature is not ideal for classifying
types recorded from volcanoes. acoustic signals, it facilitates comparison with
Please refer to Table 16.1 throughout this seismic data. As discussed in Section 16.7, seis-
section, which lists the estimated range, sound moacoustic signals in volcanoes may be coupled
or decoupled in more diverse ways than previ- The degassing burst oscillations were primar-
ously envisioned. ily in the VLP and LP bands and correlated well
with video observations of ash pulses emanat-
16.4.2â•… Kilauea volcano acoustic signals ing from the vent.
Persistent outgassing from effusive eruptions Episodic tremor at Kilauea volcano is char-
was long thought to be nearly quiescent, produ- acterized by cyclical temporal variations in
cing only faint audible hissing and white noise pressure amplitude (Fig. 16.2). The ~2–10
sounds at fumaroles and lava lakes (Richards, minute cyclic filling and draining of cavities at
1963). However, recent work has shown that Halemaumau and Pu’u ‘O’o has been observed
even low-level outgassing at Kilauea Volcano, to coincide with a “gas piston” rising and releas-
Hawaii, produces sustained infrasonic signals ing an accumulated amount of gas, as well as
that can be recorded at distances of tens of with the rising and lowering of an exposed,
kilometers (Garcés et€al., 2003; Fee and Garcés, degassing lava surface. The degassing portions
2007). Outgassing at Pu’u ‘O’o Crater, Hawaii, of this episodic tremor are consistent with ele-
is dominated by infrasonic tremor lasting from vated infrasound, while the capped or quiescent
minutes to years and is concentrated between lava surface correlates with little to no infra-
~0.4 and 10 Hz. Kilauea volcano’s infrasounds sound (Patrick et€al., 2011). The spectral content
appear to be strongly influenced by the efficient of the episodic tremor at Halemaumau is the
exsolution of the gas phase from the magma same as the aforementioned harmonic tremor,
and the interaction of this separated gas with suggesting both signals are affected by the same
the open conduits and chambers confining it. resonating cavity.
Effusion of degassed lava produces little to no
infrasound unless it encounters water or other 16.4.3â•…Strombolian and vulcanian
volatiles. acoustic signals
Kilauea volcano has also produced other types Strombolian and vulcanian explosions are some
of infrasound. Harmonic tremor from Pu’u ‘O’o of the most studied because of their relative
crater was recorded in 2007. Compressed gas abundance and availability, yet their respective
intermittently excites lava tubes, and produces waveform signatures can be difficult to differen-
a higher frequency (> 10 Hz) form of tremor tiate using purely acoustic methods. Marchetti
(Matoza et€al., 2010). Fissure eruptions also prod- et€ al. (2009) suggest that it may be possible to
uce discernable acoustic signals. During the July better discriminate between these types of
2007 fissure eruption at Kilauea’s East Rift Zone, explosive signals by using a combination of ther-
fountaining along a set of four fissures that rup- mal imaging and acoustic methods. Numerous
tured a length of ~2 km produced energetic studies at Stromboli volcano have identified
tremor with a spectral signature similar to that two primary types of explosions: short-duration
of the tremor from Pu’u ‘O’o. The signal onset explosions (~3–5 s) with relatively large ampli-
was emergent (slowly rising), suggesting no tudes (20–80 Pa at 350 m) and longer duration
explosive outburst of material occurred as the (5–15 s), more complex explosions with lower
fissures opened (Fee et€al., 2011). peak amplitudes (10–30 Pa at 350 m) (Ripepe
The ongoing 2008−2011 activity at and Marchetti, 2002; Harris and Ripepe, 2007).
Halemaumau Crater produces nearly continu- Both explosive styles occur on a periodic basis
ous harmonic infrasonic tremor in the 0.4–4 Hz and have repeatable (stable) waveforms primar-
frequency band. The amplitude and frequency ily in the LP band. Strombolian explosions are
of these spectral peaks are stable for periods of presumed to originate from overpressured gas
days, consistent with a roiling lava lake excit- slugs bursting at the surface (see Chapter 6). The
ing the overlying gas-filled cavity into resonance shorter-duration explosions have sharp, com-
(Fee et€al., 2010a). Transient degassing bursts at pressional onsets and broadband spectra with a
Halemaumau (Fee et€ al., 2010a) also suggest a peak around 5 Hz, whereas the longer-duration
shallow process that induces cavity resonance. explosions have a more complex, peaked
inferences on source durations and location the ground (Oshima and Maekawa, 2001) also
(e.g., Goerke et€ al., 1965; Tahira et€ al., 1996). A generate infrasound.
recent pilot project with arrays at 37 and 251 Very large eruptions can also produce gravity
km from Tungurahua Volcano recorded two and acoustic-gravity waves (Goerke et€al., 1965;
subplinian and one plinian eruption between Tahira et€al., 1996; Ripepe et€al., 2010) that can
2006 and 2008 (Garcés et€ al., 2008; Fee et€ al., commonly be seen as concentric cloud patterns
2010b). All three eruptions had extended dura- in satellite imagery. Acoustic-gravity waves are
tions (> 4 hours) and were characterized by affected by buoyancy, have periods longer than
high-amplitude tremor related to jetting (Section 50 s, and have unique source and propagation
16.5.3). The broadband spectrum of the jetting characteristics that are beyond the scope of this
is remarkably similar among the eruptions (Fig. chapter. Although the possibility of sustaining
16.1(b)) with the general shape resembling that VLP gravity wave oscillations in a stratified,
recorded from man-made jet engines (Matoza vesiculated magma column was proposed by
et€ al., 2009b). A 2005 phreatic eruption signal Garcés et€al. (2000), no clear evidence for these
lasting ~90 minutes at Mount St. Helens also wave types has been found to date. See Gossard
had a similar spectral signature. The transition and Hooke (1975) for an excellent introduction
of the 17 August 2006 Tungurahua eruption to gravity and acoustic-gravity waves.
from subplinian to plinian is marked by a dis-
tinct shift in the spectrum to lower frequencies
(< 0.1 Hz; Fig. 16.1). Acoustic power was found 16.5╇ ╇ Excitation
to scale roughly with eruptive intensity and ash
cloud height (Fee et€al., 2010b). The 2008 plinian In this section we discuss energetic hydro-
eruptions of Okmok and Kasatochi Volcanoes, dynamic processes that may produce measur-
Alaska, were also clearly identified with remote able pressure changes in volcanic fluids. The
infrasound arrays (Fee et€al., 2010c). most basic type of volcanic pressure disturb-
ance is an explosion. An explosion is defined as
16.4.5â•…Other types of volcano acoustic a rapid expansion of matter into a volume much
signals greater than its original volume (Cooper and
LP seismic events have been attributed to hydro- Kurowski, 1996). Explosions are represented as
thermal and magmatic activity in shallow vol- transients with a positive pressure onset (com-
canic cracks and conduits (Chouet, 1996). These pression), defined by its peak pressure and rise
events have also produced notable acoustic coun- time, and followed by a negative pressure (rar-
terparts (Yamasato, 1998; Matoza et€ al., 2009a). efaction) corresponding to gas overexpansion.
Infrasonic LPs at Mount St. Helens have impul- Such bipolar explosion pulses may be described
sive onsets, durations of 5−10 s, and relatively by the duration, rise time, and peak pressure
flat, broadband spectra. These events are highly of the signal (Fig. 16.3, ANSI S12.7–1986, 1986).
regular in their occurrence and have stable We note that in the volcano-acoustics litera-
waveform features, indicating a non-destructive, ture, volcanic explosion waveforms are some-
repetitive source (Matoza et€al., 2009a). Acoustic times incorrectly referred to as N-waves. True
records of pyroclastic density currents (PDCs) N-waves, such as those from supersonic sources
are rare or commonly masked by concurrent (e.g., Garcés et€ al., 2004) and the detonation of
jetting signals. Yamasato (1997) detected PDCs high explosives (Gossard and Hooke, 1975), have
acoustically at Unzen Volcano, Japan. and was an unambiguous N-shape. More continuous gas-
able to track their progression using arrival release processes, such as subplinian and plin-
times and Doppler shift. Ripepe et€al. (2010) also ian eruptions, can last for several hours and yet
tracked PDCs with a single infrasound array. The are sometimes referred to as explosive. In vol-
partial collapse of dacitic lava domes (Green and canology, explosive activity refers to eruptions
Neuberg, 2005; Moran et€ al., 2008a) and explo- “where magma is torn apart and ejected from
sive blowout of gas-charged blocks impacting the vent as clots or blobs within a stream of gas”
per second. Thus highly vesiculated MGMs may interpreted within the MGM and BLEVE frame-
be vulnerable to lower fragmentation thresh- work. Explosion waveforms from magmatic
olds (< 70% void fraction) as relatively low flow activity at Tungurahua volcano (Fig. 16.4) may
speeds reach near supersonic conditions and represent the sudden release of gas contained
blow these foams apart, triggering an explosive at extremely high pressure in the magma (i.e.,
eruption (Garcés, 2000). an impulsive fragmentation), whereas the more
sustained signals at Tungurahua associated with
16.5.2â•…Steam boilers and popping jetting activity (Fig. 16.5(a)) can be understood
bubbles in terms of runaway fragmentation tapping a
The depressurization of a large static or greater MGM volume that persists for hours.
pressure-compensated gas pocket that has From the vantage point of volcano acoustics,
already separated from the melt is easy to visu- the aim is to quantitatively relate the recorded
alize. The overpressure in such a system may be sounds to fluid dynamic processes resulting
estimated as the sum of the lithostatic overpres- from unstable MGMs.
sure, ΔPl, and the rock or lava yield stress. The Figure 16.5(b) shows the infrasonic spec-
yield stress is the force per unit area beyond trogram recorded at a range of 37 km from
which permanent deformation will occur, and Tungurahua volcano during the sustained sub-
may have values on the order of 10–1000 MPa, plinian to plinian eruption of 17 August 2006
with higher values possible for plastic materi- (PSD in Fig. 16.1). Figure 16.5(c) shows the
als. Lithostatic overpressure can be readily esti- estimated acoustic power above 0.1 Hz (with
mated from: a peak of ~30 MW) and the associated ash
cloud heights derived from satellite imagery
ΔPl = ρr g Δz ~ (2500 kg m−3)(9.8 m s−2)
(Fee et€al., 2010b). Although there are multiple
Δz ~ Δz/40 MPa m−1, (16.13)
discrete impulsive explosion waveforms (Fig.
where Δz is depth in meters and the density ρr 16.5(a)), the dominant signal feature is the sus-
is representative of somewhat fractured and/or tained broadband, high-amplitude infrasonic
porous crust material. A gas pocket will remain tremor. Matoza et€ al. (2009b) proposed that
in equilibrium with the ambient pressure or these long-duration infrasonic signals were pro-
containment (skin) pressure. Because the ambi- duced by turbulent jetting of ash−gas mixtures
ent atmospheric pressure is ~0.1 MPa, it is pos- in large-scale eruption jets (>300 m diameter at
sible to overpressurize gas with a sealed layer Tungurahua) and represent an infrasonic form
of rock or a magma film. Depending on the of jet noise. Jet noise is produced at audible fre-
containment rupture mechanism, the pressure quencies by man-made aircraft and rockets and
may be released as a jet, a small explosive burst results from the turbulent flow of air out of a jet
followed by a jet, or a burst where all the gas is engine. The acoustic frequencies at which jet
released simultaneously (Birk and Cunningham, noise is radiated scale with the Strouhal num-
1994). If the gas is already exsolved, as in the ber (discussed further in Section 16.7.5) and are
case of a static or ascending bubble, this process related to the length scales of the jet flow (in
will only last for as long as it takes to release the particular, the diameter). The jetting activity
gas volume. As demonstrated by Ichihara et€al. at the lower portion of eruption columns is of
(2009), a shallow but submerged explosion trig- large diameter compared to typical man-made
gered by sudden depressurization of an MGM jets; thus, the acoustic frequencies of radiation
would appear visually similar to a popping sur- are much lower than those observed from jet
face bubble, and may be similar in its source engines (Fig. 16.1a). Matoza et€al. (2009b) used
pressure signature. the expanded jet diameter as the characteris-
tic length scale (Eq. (16.28), Section 16.7.5), and
16.5.3â•… Transients and jets found the Strouhal number ranged from 0.06
Many of the transient and continuous acous- to 0.4 for the volcanic jets considered in his
tic signals presented in Section 16.4 can be study.
Figure 16.4╇╇Time-corrected
Tungurahua (a) explosion
waveforms and (b) spectra
recorded at a distance of 5
km, 37 km, and 251 km. As the
distance is increased, the signal
loses high-frequency energy
and its duration is extended.
Nearest station data courtesy
of H. Kumagai. See color plates
section.
2πds
16.6╇╇ Acoustic radiation << 1, (16.14)
λ
The expressions presented in this section may where λ is the acoustic wavelength. At a fre-
be regarded as extremely useful fictions, applic- quency of 1 Hz for air at 20 °C, this corresponds
able to ideal linearized cases where complex to ds << 50 m. At steam temperatures of 700 °C
processes are represented as simple sources that may be encountered inside magma con-
or combinations of simple sources. The pri- duits, this condition corresponds to ds << 100 m.
mary justification for this approach is that it Although this compactness condition may not
has performed surprisingly well in controlled always be met, it provides a tractable starting
acoustic environments, so we extend it to the point for discussing sources of volcanic sound.
Figure 16.5╇╇ (a) Waveforms, (b) spectrogram, and (c) if a fluid of density ρ0 flows with a speed u in a
radiated acoustic power (in black) and ash cloud top (in section of cross-sectional area A with a time his-
green) for the suplinian to plinian Tungurahua eruption of 17 tory given by:
August 2006. The eruption was characterized by continuous
u (t) = u0 + u s (t), (16.16a)
jetting for >10 hours, with some explosions as well. Acoustic
power broadly scales with ash cloud height for this eruption,
where u0 is the steady flow speed, and us(t) is
particularly during the plinian phase of the eruption around
17 August 0600 UTC. From Fee et€al. (2010b). See color the unsteady speed as a function of time, the
plates section. total rate of mass outflow m(t) can be estimated
from:
Lighthill (1978) provides an excellent discus- m ( t ) = ρ0 Au ( t ) = ρ0 Au0 + ρ0 Au s ( t ) , (16.16b)
sion on the virtues of compactness. In particular,
for a given mass outflow rate m(t) in an infinite and hence its rate of change is
homogeneous volume, the acoustic pressure p(t)
at a distance r from the source is given by: dm(t) du (t)
= ρ0 A s . (16.16c)
dt dt
1 dm(t ) ρ d 2 V (t )
p (t ) = = 0 , (16.15)
4 π r dt t − r / c 4 π r dt 2 t − r / c Note that Eqs. (16.15) and (16.16c) show that
an increase in pressure is associated with fluid
where V is the volume displaced by the mass acceleration. The first term in Eq. (16.16b) is the
of a fluid with equilibrium density ρ0. A severe steady mass flux ρ0Au0 which is routinely used
limitation on estimating mass flux from acous- in incompressible fluid dynamics and is gener-
tic measurements is that slowly fluctuating ally the dominant mass contribution. Because
(nearly steady-state) mass-flux contributions do acoustic measurements are by definition per-
not contribute to the sound field. For example, turbations from the steady pressure field, Eq.
(16.16c) shows that the time-invariant steady source would be placed in an anechoic cham-
mass flux will not produce sound. A common ber and surrounded by a grid of microphones
assumption in compressible fluid flows is that to characterize its frequency-dependent radi-
the time-varying unsteady component us(t) is a ation pattern. Such rigor is difficult in volcanic
fraction of the steady stream velocity u0, so that environments.
it is possible to estimate the stream velocity Let us define Pm as the source pressure at 1 m
from the perturbation velocity. Estimates of radiated by a monopole, represented by a pulsat-
the steady mass flux term are sensitive to the ing sphere in free space. The free-space mono-
details of the models used to represent the fluid pole is one of the most useful approximations in
dynamics of a source process, and may not be the field of acoustics, as it acts as a simple source
robust because of the large number of free vari- and is used as a reference for all other compact
ables (Garcés, 1997). source representations. Of slightly more prac-
Exact expressions for acoustic sources can tical value is the expression for a source on or
generally be derived for only the simplest geom- near Earth’s surface, which is modeled as a baf-
etries and propagation conditions. Most solu- fled source. A baffle is an acoustic term that usu-
tions are rich in complex detail and structure ally refers to the flat plate to which a speaker is
near the source, but stabilize substantially far mounted, but here is approximated as a solid
from the source. In addition, many mathemat- plane bounding a half-space. The solution for
ical expressions have asymptotic expansions the pressure radiated by a hemispherical source
that simplify at a distance of a few wavelengths in a baffle is identical to that of the monopole,
from the source as the waves lose their curva- because both the environment and the source
ture and start resembling plane waves. The typ- strength are halved. Bursting bubbles described
ical criterion for the far-field is: in Section 16.5.2 are examples of hemispherical
sources, as the gas overpressure would be com-
2π r c 54 pletely released to the atmosphere€– the path of
>> 1 ⇒ r >> >> . (16.17)
λ 2π f f least resistance€– with little or no acoustic radi-
ation into the melt or ground.
At 0.02, 0.1, 1, and 10 Hz, the far-field condition In contrast, the far-field solution for a spher-
corresponds to r >> 3 km, 500 m, 50 m, and 5 m, ical source close to an infinite baffle and radiat-
respectively. ing into a half space is given by:
Moran et€ al. (2008a) exploited the compact-
ness condition (Eq. (16.14)) to model a VLP acous- 2Pm
Ps / 2 = . (16.18)
tic signal resulting from a large rockfall from r
the 2004–08 Mount St. Helens lava extrusion.
Similar modeling approaches have been applied Sound from a compact industrial source placed
to strombolian bubble bursts at Mount Erebus, on a stand is commonly represented by this
Antarctica (Johnson et€al., 2008). equation. A classic dipole may be represented
as two monopoles pulsating in opposite phase
16.6.2â•…Multipole expansions for surface to one another. Such a condition, which is typ-
and subaerial sources ical of reflection of underwater sound from a
An acoustic source may be represented by shallow source near the ocean−atmosphere
distributions of simple sources, an approach interface, is useful for visualizing the inter-
known as multipole expansion (Morse and Ingard, ference pattern caused by reflection along an
1968). Multipoles permit a ready visualiza- infinite plane (the Lloyd’s mirror effect) but is
tion of the intriguing interference patterns an inadequate way to model reflections in com-
produced by the superposition of multiple plex environments. Acoustic reflection from
sources. Multipole representations may be multiple irregular walls is better treated as a
practical in industrial environments where one reverberation problem, to be discussed in more
may control recording conditions. Ideally, the detail in Section 16.7.
Note that far from the source θ ~ π/2 and ka sinθ at a distance are imprinted with the resonance
~ 2πa/λ. If, in addition, the wavelength is much and attenuation characteristics of the volume.
greater than the piston radius so that λ << a, This allows for modeling of seismic and acoustic
then waveforms to infer the geometry and fluid com-
position of volcanic resonators. A critical step
forward was made in volcano seismology when
lim ∏ p = ∏ h . (16.23b)
ka → 0
it was postulated that acoustic resonance effects
are important in generating observed seismicity
If the spatial acoustic source distribution is
at volcanoes (e.g., Aki et€ al., 1977). Since then
unknown, or if the recorded field is a super-
much work has been done in volcano seismology
position of subsurface, surface, and airborne
concerning solid−fluid interface waves or crack
sources (for example, a transient source in a
waves that occur from wave propagation in small
deep crater accompanied by vigorous jetting) it
volumes of fluid trapped in a solid (Chapters
would be judicious to use Eq. (16.20) for the free
15). Decades ago, volcano seismology identified
space solution, as power estimates would only
tremor and long-period events as signals that may
be a factor of 2 (3 dB) different from Eq. (16.21)
be indicative of resonant volcanic structures. In
or (16.22). For explosions and other transient
contrast, the first infrasonic measurements of har-
events, the total energy in joules for the event
monic tremor were made at Sakurajima in 1996
may be estimated by E = Ts Π, where care must
(Sakai et€al., 1996). In the past decade, because of
be taken that the time of overpressure integra-
an increase in the quantity and quality of acoustic
tion, Ts, only corresponds to the main pulse. As
measurements, numerous observations of infra-
an example, one of the Halemaumau vent (HV)
sonic harmonic tremor have been reported from
VLP degassing bursts had an estimated energy of
volcanoes worldwide (Section 16.4). This section
~106 J, an order of magnitude lower than the HV
focuses on resonance effects in volcanoes. We
ensuing tremor energy radiated in one hour (~3
examine several processes in a volcanic fluid sys-
kW = 107 J hr−1; Fee et€al., 2010a). If the source
tem that could result in sharply defined spectral
is within the erupting vent or more than a few
peaks in acoustic data. In particular, we discuss
tens of kilometers from the recording station,
acoustic resonance or reverberation (standing
other propagation effects need to be considered
waves), volume (Helmholtz) resonance, and
as well (Fee and Garcés, 2007). The effect of the
flow-induced oscillations. We note that repeti-
atmosphere on telesonic (> 250 km) propagation
tive evenlyspaced impulsive events (e.g., Powell
is beyond the scope of this chapter, and is cov-
and Neuberg, 2003) may also produce harmonic
ered in greater detail in Le Pichon et€al. (2010).
spectra without resonance. The combination of
repetitive pulses and resonance was considered
by Garcés and McNutt (1997).
16.7╇ ╇ Resonant oscillations
16.7.1â•… Acoustic resonance
Resonance is an unavoidable propagation effect Acoustic resonance consists of constructively
in most volcanic fluid systems. Magma ascends to interfering echoes, and is the natural response
Earth’s surface through a complicated network of of bounded fluids. Acoustic resonance is a famil-
cracks, chambers, and conduits, and hydrother- iar concept widely exploited in the design of
mal systems and gas accumulations can be stored musical instruments to produce harmonious
in subsurface reservoirs or near-surface cavities. tones (e.g., Fletcher and Rossing, 1998). Solid
If pressure disturbances occur in any of these pipe sections of intricate geometry contain-
confined fluid volumes, acoustic energy is par- ing air columns (e.g., an organ pipe) produce
tially trapped and the volumes resonate at their sound with predictable spectra. By varying the
natural frequencies, or eigenfrequencies. As the effective length of a resonator, different nat-
elastic energy leaks away from these regions, the ural sound frequencies can be produced, which
resulting seismic and acoustic signals observed are perceived as changes in pitch. The sound
radiated by an instrument is a complex super- pipe modes of Eqs. (16.24a) and (16.24b)) are
position of the different resonant oscillations of given by:
the instrument body. Unfortunately, in volcanic 2
conduits and cracks, the fluid composition in cj
f s, m = 1s + f m2 , (16.25)
the resonator can be highly variable and largely 2π a
unknown. Consider a one-dimensional oscilla-
where fs,m is the radial resonance associated
tor such as the classic organ pipe modes, with
with the sth zero j1s of the Bessel function J1
eigenfrequencies given by:
and the mth longitudinal mode, c is the sound
1 speed in the magma, a is the conduit radius,
m − c (16.24a) and L is the conduit length. In general, two-
2
fm = , where m = 1, 2,...
2L and three-dimensional standing wave patterns
are produced inside confined volumes, lead-
or ing to complex and possibly degenerate mode
sequences that depend on the geometry and fluid
mc
fm = , where m = 1, 2, 3 ... (16.24b) properties of the resonator. One of the important
2L results of Buckingham and Garcés (1996) is that
the higher-order modes, which correspond to
where fm are the resonant modes, c is the sound
higher frequencies, radiate nearly vertically and
speed in the resonator, L is the effective length
are not efficiently propagated into the far-field.
of the resonator, m is the mode number, and the
The predicted radiation pattern is reminiscent to
form of equation depends on the whether the
that of a piston in a baffle, with monopole radi-
pipe is closed and open at one end (Eq. 16.24a)
ation dominant when the wavelength is larger
or open at both ends (Eq. 16.24b). The first mode,
than the vent aperture, and a strong vertically
m = 1, is generally dominant and is known as
directed radiation pattern at high frequencies.
the fundamental, with all higher modes called
This helps to explain why explosive sources in
overtones. Whereas in musical instruments the
vents and craters rapidly lose audible frequen-
resonating fluid is air and c is well known, in a
cies with distance. Furthermore, liquid−gas
volcano the fluid may be magma, gas, a dusty
interfaces are inefficient at transmitting sound,
ash−gas mixture, juvenile or meteoric water,
and act as good reflecting boundaries. However,
mud, or a multiphase combination of the above.
sources submerged within magmatic fluids may
The in situ acoustic properties of these volcanic
also radiate more efficiently than anticipated
fluids are critical parameters defining the reson-
because of the recently postulated anomalous
ant properties of volcanic conduits and cracks,
transparency of infrasound at the liquid−air inter-
yet these properties can only be crudely esti-
face (Godin, 2006), which to date has not been
mated from geophysical and geochemical meth-
incorporated into volcano-acoustic models.
ods. Nevertheless, a nondimensional approach
The formulation of Garcés (2000) addresses
enables inferences about volcanic resonators
the resonant properties of a variable-width tube
that are not tied to fixed values of key geophys-
of fluid that may be moving at high velocity
ical properties of the fluid.
relative to the sound speed of the flow. To a
Buckingham and Garcés (1996) developed
first-order approximation, the spacing between
a canonical model for the atmospheric sound
frequency peaks is given by:
field generated by a resonant magma conduit
excited by an explosive compact source. To
c (1 − M 2 )
obtain a tractable analytic solution, a num- ∆f = , (16.26)
2L
ber of simplifying assumptions were made. A
repetition of this analytic solution is beyond where c is the sound speed of the material in
the scope of this chapter. However, we remark the conduit, M = U/c is the Mach number of the
that the symmetric radial modes of a cylin- flow, and L is the effective length of the conduit.
drical conduit (in addition to the longitudinal Hence, the eigenfrequency variations known
as gliding (Garcés and Le Pichon, 2009) could be sound source by acting as an echo chamber. In
explained by either a change in the effective rectangular coordinates, the natural oscillation
length of the conduit L, or changes in the flow frequencies of a rectangular stiff-walled cham-
velocity U or sound speed c. Both sound speed ber are given by:
and flow velocity vary readily with a change in
2 2 2
the mean vesicularity of a magma−gas mixture. c lπ mπ nπ
f lmn = + + , (16.28)
2π Lz
Lx L y
16.7.2â•… Helmholtz resonance
For acoustic wavelengths greater than any of where l, m, n = 0, 1, 2, … are the mode numbers,
the linear dimensions of a volume, Helmholtz and Lx, Ly, and Lz are the x, y, and z longitudinal
resonance can also be induced. A Helmholtz dimensions of the chamber. The eigenfrequen-
resonator is a stiff-walled cavity connected to cies in Eq. (16.28) can be considered as travel-
the atmosphere through an opening with a ing plane waves in their respective directions
neck. The cavity and opening would represent (Kinsler et€al., 1982).
a vented volcanic chamber. When fluid is We can rewrite Eq. (16.28) as:
pushed out of the cavity, low pressure is cre-
ated within the cavity, which responds by pull- 2
L 2
2
2 Lz f lmn L
ing the air back in, and vice versa. The system f lmn = = z l2 + z m + n 2 . (16.29)
c f 1z Ly
may then sustain oscillations at a frequency
Lx
given by:
If the vertical length Lz and the volume V of the
c Sa chamber are constrained, we can specify a cross-
fH = , (16.27) sectional area A = LxLy and leave only one free
2π LH V
variable Lx
where Sa is the cross-sectional area of the neck,
2 2
LH is the effective length of the neck, c is the f lmn L L L
= z l 2 + z x m 2 + n 2 , (16.30)
sound speed of the fluid, and V is the volume f1z L
x A
of the cavity. If there is no neck, such as in the
case of a vent within a thin lava roof, the effect- and now it is possible to plot the normalized
ive neck length LH may be approximated by 1.7a, eigenfrequency as a function of only one dimen-
where a is the radius of the neck opening with sion, Lx. Fee et€ al. (2010a) performed such an
area Sa = πa2 (Kinsler et€al., 1982). eigenfrequency analysis assuming Helmholtz res-
The dimensional requisites for Helmholtz onance and using LIDAR measurements to con-
resonance are compatible with the observed strain the volume and depth of Halemaumau’s
geometry of some near-surface volcanic cavities gas-filled chamber (0.5 Hz peak in Fig. 16.1(b)),
(Fee et€al., 2010a). A given cavity volume can sus- and obtained a reasonably good match between
tain a lower fundamental oscillation frequency theory and observation. Decades of architectural
via Helmholtz resonance than it could produce acoustic studies attest to the difficulties of deriv-
by the longitudinal modes in Eq. (16.24). ing precise mode amplitude solutions for rooms
with complex shapes. Volcanic chambers would
16.7.3â•… Echo chambers certainly deviate from a perfect rectangular vol-
Although a Helmholtz resonator is predicted to ume, and the observed mode amplitudes would
oscillate at a single low frequency, resonance depend on the source position, the walls’ exact
peaks at higher frequencies are possible within a shape, texture, and sound absorption properties,
cavity in the form of three-dimensional standing and the radiation condition at the vent. However,
acoustic waves analogous to the one-dimensional all these effects are likely to be more aggravated
pipe modes described in Section 16.7.1. The sim- as the frequency increases and the acoustic wave-
plest three-dimensional cavity is a rectangular length becomes comparable to the spatial scales of
volume, which would add reverberation to the the irregularities in the chamber. The good match
between theory and observation at low frequen- This disturbance can interact with the shear
cies obtained by Fee et€al. (2010a) is encouraging. layer upstream. Shear layers are very sensitive
to minor changes in the pressure conditions
16.7.4â•… Flow-induced oscillations around them, such that tiny acoustic pressure
In previous sections, we discussed how a volcanic oscillations resulting from interaction with
cavity may reverberate to produce the sharp- the solid object can result in the generation
lypeaked tones observed in infrasonic harmonic of vortices in the shear layer. These vortices
tremor (Section 16.4). In addition, in Section 16.5 are then carried back downstream in the flow
we introduced various means of exciting resona- where they again reach the solid object. The
tors (the s(t) function in Eq. (16.12)). Thus we have interaction between the upstream propagat-
treated a resonant volcanic system as separable, ing acoustic or hydrodynamic disturbance and
in which an impulsive or sustained broadband the downstream propagating vortices results
driving function can excite a cavity into produ- in a closed feedback loop, which is strongest
cing transient resonant signals (LPs) or harmonic at particular frequencies. When examining
tremor. However, flow-induced oscillations in a flow-induced oscillations, frequencies are usu-
volume also provide nonlinear mechanisms that ally non-dimensionalized in the tonal Strouhal
can yield lower natural frequencies than pre- number, given by:
dicted from purely acoustic means.
In musical instruments, the resonant prop- fL
St = , (16.31)
erties of the pressure or force-function driving U
cavity resonance can be critical. For example,
where L is the length scale of the process (m),
flute tones are a result of flow-induced edge tones
and U is the jet flow velocity (m s−1). Rossiter
generated by the jet of air from the player’s lips
(1964) found the Strouhal numbers of these
onto an edge in the mouthpiece, and the amp-
processes agree with the empirical equation:
lification of these tones via the resonator body
(the pipe section) (Fletcher and Rossing, 1998). m −γ
St = ,
When examined in more detail, the relationship 1 (16.32)
M+
between the driving mechanism and resonator K
response is nonlinear, and involves a complex
feedback process between the acoustic modes where m is the mode number (m = 1, 2, 3, …),
of the resonator body and the air jet exciting M€ = U/c is the Mach number, and γ and K are
the modes. This process leads to mode-locking in empirical constants. Rossiter (1964) then
which particular tones are preferentially ampli- deduced that this can be expressed in terms of
fied (Fletcher, 1999). Such complex acoustic feed- the physical parameters of the system as:
back processes may provide an explanation for L L (m − γ )
infrasonic harmonic tremor (Matoza et€al., 2010). + = , (16.33)
Uc c fm
The edge tone is just one example of a fam-
ily of processes in which tones are created by which is known as Rossiter’s equation (Howe,
the interaction of shear layers with solid bound- 1998). Here, Uc is the velocity of propagation
aries (Rockwell and Naudascher, 1979). Other of the vortices, c is the sound speed, fm are the
examples include the hole-tone, which is pro- Rossiter modes and γ represents a phase lag. The
duced by flow of an axisymmetric jet from one constant K in Eq. (16.32) is the ratio of Uc to U,
plate impinging on a second plate with a hole i.e., Uc = KU. It remains an empirical constant but
in it, and the flow of air over a cavity (Rossiter, is ~0.4–0.6 for most processes (Rossiter, 1964).
1964). Each is believed to result from a similar The coupling of Rossiter modes of subsonic
feedback process. Once the flow encounters the volcanic jet flows to volcanic resonator bod-
solid object downstream, a disturbance is cre- ies such as conduits, cracks, and near-surface
ated that can propagate back upstream either cavities remains an open and exciting area for
as an acoustic or hydrodynamic disturbance. future study.
Encyclopedia of Complexity and Systems Science, ed. Journal of Geophysical Research, 114, B03210,
R.╯A. Meyers. Springer. doi:10.1029/2008JB005792.
Garcés, M.╯A., Hagerty, M.╯T. and Schwartz, S.╯Y. Johnson, J.╯B. (2003). Generation and propagation
(1998). Magma acoustics and time-varying of infrasonic airwaves from volcanic explosions.
melt properties at Arenal Volcano, Costa Rica. Journal of Volcanology and Geothermal Research, 121,
Geophysical Research Letters, 25(13), 2293–2296. 1–14, doi:10.1016/ S0377–0273(02)00408–0.
Garcés, M.╯A., McNutt, S.╯R., Hansen, R.╯A. and Johnson, J.╯B. and Lees, J.╯M. (2000). Plugs and chugs;
Eichelberger, J.╯C. (2000). Application of seismic and acoustic observations of degassing
wave-theoretical seismoacoustic models to the explosions at Karymsky, Russia and Sangay,
interpretation of explosion and eruption tremor Ecuador. Journal of Volcanology and Geothermal
signals radiated by Pavlof Volcano, Alaska. Journal Research, 101, 67–82.
of Geophysical Research, 105, 3039–3058. Johnson, J.╯B. and Ripepe, M. (2011). Volcano
Garcés, M., Harris, A., Hetzer, C. et€al. (2003). infrasound: A review. Journal of Volcanology and
Infrasonic tremor observed at Kilauea Volcano, Geothermal Research, in press.
Hawai’i. Geophysical Research Letters, 30, Johnson, J.╯B., Aster, R.╯C. and Kyle, P.╯R. (2004).
doi:10.1029/2003GL018038. Volcanic eruptions observed with infrasound.
Garcés, M., Bass, H., Drob, D. et€al. (2004). Forensic Geophysical Research Letters, 31, L14604,
studies of infrasound from massive hypersonic doi:10.1029/2004GL020020.
sources. Eos, Transactions of the American Geophysical Johnson, J., Aster, R., Jones, K.╯R., Kyle, P.
Union, 85(43), 443. and Mcintosh, B. (2008). Acoustic source
Garcés, M., Fee, D., Steffke, A. et€al. (2008). Capturing characterization of impulsive Strombolian
the acoustic fingerprint of stratospheric ash eruptions from the Mount Erebus lava lake. Journal
injection. Eos, Transactions of the American Geophysical of Volcanology and Geothermal Research, 177(3),
Union, 89(40), 377. 673–686.
Godin, O.╯A. (2006). Anomalous transparency of Kinsler, L.╯E., Frey, A.╯R., Coppens, A.╯B. and Sanders,
water-air interface for low-frequency sound. J.╯V. (1982). Fundamentals of Acoustics, Third Edition.
Physical Review Letters, 97, 164301. Wiley.
Goerke, V.╯H., Young, J.╯M. and Cook, R.╯K. (1965). Le Pichon, A., Blanc, E. and Hauchecorne, A.
Infrasonic observations of the May 16, 1963, (2010). Infrasound Monitoring for Atmospheric Studies.
volcanic explosion on the island of Bali. Journal of Springer.
Geophysical Research, 70(24), 6017–6022. Leventhall, H.╯G., Benton, S. and Pelmear, P. (2003). A
Gossard, E.╯E. and Hooke, W.╯H. (1975). Waves in Review of Published Research on Low Frequency Noise and
the Atmosphere: Atmospheric Infrasound and Gravity its Effects. Report for Department of Environment,
Waves€– Their Generation and Propagation. Elsevier. Food, and Rural Affairs, UK.
Green, D.╯N. and Neuberg, J. (2005). Seismic and Lighthill, M.╯J. (1954). On sound generated
infrasonic signals associated with an unusual aerodynamically. II. Turbulence as a source of
collapse event at the Soufriere Hills volcano, sound. Proceeedings of the Royal Society of London, A
Montserrat. Geophysical Research Letters, 32, L07308, 222, 1–32.
doi:10.1029/2004GL022265. Lighthill, M.╯J. (1978). Waves in Fluids. Cambridge
Harris, A. and Ripepe, M. (2007). Synergy of multiple University Press.
geophysical approaches to unravel explosive Marchetti, E., Ripepe, M., Harris, A.╯J. L. and Delle
eruption conduit and source dynamics: A case Donne, D. (2009). Tracing the differences between
study from Stromboli. Chemie Der Erde, 67(1), 1–35. Vulcanian and Strombolian explosions using
Hedlin, M.╯A. H. and Raspet, R. (2003). Infrasonic infrasonic and thermal radiation energy. Earth and
wind-noise reduction by barriers and spatial filters. Planetary Science Letters, 279(3–4), 273–281.
Journal of the Acoustical Society of America, 114(3), Matoza, R.╯S., Garcés, M.╯A., Chouet, B.╯A. et€al.
1379–1386. (2009a). The source of infrasound associated
Howe, M.╯S. (1998). Acoustics of Fluid–Structure with long-period events at Mount St. Helens.
Interactions. Cambridge University Press, 560 pp. Journal of Geophysical Research, 114, B04305,
Ichihara, M., Ripepe, M., Goto, A. et€al. (2009). doi:10.1029/2008JB006128.
Airwaves generated by an underwater Matoza, R.╯S., Fee, D., Garcés, M.╯A. et€al. (2009b).
explosion: Implications for volcanic infrasound. Infrasonic jet noise from volcanic eruptions.
Geophysical Research Letters, 36, L08303, Ponceau, D. and Bosca, L. (2010). Low-noise
doi:10.1029/2008GL036486. broadband microbarometers In Infrasound
Matoza, R.╯S., Fee, D. and Garcés, M. (2010). Monitoring for Atmospheric Studies, ed. A. Le Pichon.
Infrasonic tremor wavefield of the Pu‘u O‘o crater Netherlands: Springer, doi:10.1007/978–1-4020–
complex and lava tube system, Hawaii, in April 9508–5_24, pp. 119–140.
2007. Journal of Geophysical Research, 115, B12312, Powell, T.╯W. and Neuberg, J. (2003). Time dependent
doi:10.1029/2009JB007192. features in tremor spectra. Journal of Volcanology and
Moran, S.╯C., Matoza, R.╯S., Garcés, M.╯A. et€al. Geothermal Research, 128(1–3), 177–185.
(2008a). Seismic and acoustic recordings of an Richards, A.╯F. (1963). Volcanic sounds, investigation
unusually large rockfall at Mount St. Helens, and analysis. Journal of Geophysical Research, 68,
Washington. Geophysical Research Letters, 35, L19302, 919–928.
doi:10.1029/2008GL035176. Ripepe, M. and Marchetti, E. (2002). Array
Moran, S.╯C., Moran, S.╯C., Mcchesney, P.╯J. and tracking of infrasonic sources at Stromboli
Lockhart, A.╯B. (2008b). Seismicity and infrasound volcano. Geophysical Research Letters, 29(22), 2076,
associated with explosions at Mount St. Helens, doi:10.1029/2002GL015452.
2004–2005. In A Volcano Rekindled: The First Year of Ripepe, M., De Angelis, S., Lacanna, G. and
Renewed Eruptions at Mount St. Helens, 2004–2006, ed. Voight, B. (2010). Observation of infrasonic
D.╯R. Sherrod, W.╯E. Scott and P.╯H. Stauffer. U.S. and gravity waves at Soufrière Hills Volcano,
Geological Survey Professional Paper 1750, pp. Montserrat. Geophysical Research Letters, 37, L00E14,
111–127. doi:10.1029/2010GL042557.
Morrissey, M.╯M. and Mastin, L.╯G. (2000). Vulcanian Rockwell, D. and Naudascher, E. (1979).
eruptions. In Encyclopedia of Volcanoes, ed. H. Self-sustained oscillations of impinging free
Sigurdsson. San Diego, CA: Academic Press, San shear layers. Annual Review of Fluid Mechanics, 11,
Diego, pp. 463–475. 67–94.
Morse, P.╯M. and Ingard, K.╯U. (1968). Theoretical Rossiter, J.╯E. (1964). Wind-tunnel experiments on
Acoustics. Princeton, NJ: Princeton University Press. the flow over rectangular cavities at subsonic
Olson, J.╯V. and Szuberla, C.╯A. L. (2009). Processing and transonic speeds. Aeronautical Research Council
infrasonic array data. In Handbook of Signal Processing Reports and Memoranda 3438. London: HMSO.
in Acoustics, ed. D. Havelock, S. Kuwano and M. Sakai, T., Yamasato, H. and Uhira, K. (1996).
Vorlander. New York: Springer, pp. 1487–1496. Infrasound accompanying C-type tremor at
Oppenheim, A.╯V. and Schafer, R.╯W. (1989). Sakurajima volcano. Bulletin of the Volcanological
Discrete-Time Signal Processing. Prentice Hall. Society of Japan, 41, 181–185 (in Japanese).
Oshima, H. and Maekawa, T. (2001). Excitation Strachey, R. (1888). On the air waves and sounds
process of infrasonic waves associated with caused by the eruption of Krakatoa in August,
Merapi-type pyroclastic flow as revealed by a new 1883. In Krakatau 1883 (published 1983), ed. T. Simkin
recording system. Geophysical Research Letters, 28(6), and R.╯S. Fiske. Smithsonian Institution Press, pp.
1099–1102. 368–374.
Parfitt, E. and Wilson, L. (2008). Fundamentals of Tahira, M., Nomura, M., Sawada, Y. and Kamo, K.
Physical Volcanology. Wiley, 256 pp. (1996). Infrasonic and acoustic-gravity waves
Patrick, M.╯R., Wilson, D., Fee, D., Orr, T. and generated by the Mount Pinatubo eruption of
Swanson, D. (2011). Shallow degassing events as a June 15, 1991. In Fire and Mud: Eruptions and Lahars
trigger for very-long-period seismicity at Kılauea of Mount Pinatubo, Philippines, ed. C.╯G. Newhall
Volcano, Hawai‘i. Bulletin of Volcanology, doi:10.1007/ and R.╯S. Punongbayan. Seattle, WA: University of
s00445–011–0475-y. Washington Press, pp. 601–614.
Petersen, T., De Angelis, S., Tytgat, G. and McNutt, Vergniolle, S. and Brandeis, G. (1996). Strombolian
S.╯R. (2006). Local infrasound observations of large explosions. 1. A large bubble breaking at
ash explosions at Augustine Volcano, Alaska, the surface of a lava column as a source of
during January 11–28, 2006. Geophysical Research sound. Journal of Geophysical Research, 101(B9),
Letters, 33, L12303, doi:10.1029/2006GL026491. 20,433–20,447.
Pierce, A.╯D. (1981). Acoustics: An introduction to Wood, A.╯B. (1964). A Textbook of Sound. London: Bell.
Its Physical Principles and Applications. New York: Woulff, G. and Mcgetchin, T.╯R. (1976). Acoustic
McGraw-Hill. noise from volcanoes: Theory and experiment.
Geophysical Journal of the Royal Astronomical Society, Use the mean value theorem to derive the
45, 601–616. SPL between 1–2 Hz from the spectra shown
Yamasato, H. (1997). Quantitative analysis of in Figure 16.1, and compare to the results in
pyroclastic flows using infrasonic and seismic data Table 16.1.
at Unzen volcano, Japan. Journal of Physics of the 16.2╇Estimate the SPL for the Halemaumau Vent
Earth, 45(6), 397–416. between 0.3 and 0.6 Hz, and the radiated
Yamasato, H. (1998). Nature of infrasonic pulse source power in watts.
accompanying low frequency earthquake at Unzen 16.3╇If a volcanic jet has an expanded diameter of
Volcano, Japan. Bulletin of the Volcanological Society of 10 m and peak frequency of 1 Hz, what range
Japan, 43, 1–13. of jet velocities would yield a Strouhal number
range of 0.06–0.4?
Planetary volcanism
Rosaly M.╯C. Lopes, Sarah A. Fagents, Karl L. Mitchell
and Tracy K.╯P. Gregg
Modeling Volcanic Processes: The Physics and Mathematics of Volcanism, eds. Sarah A. Fagents, Tracy K.╯P. Gregg, and Rosaly M.╯C.
Lopes. Published by Cambridge University Press. © Cambridge University Press 2012.
volcanic landforms, and for magma chemistry eruptive products, are affected by the ambi-
(Section 17.2.4). Despite these differences, the ent conditions on different bodies (Table 17.1).
eruption styles and products on other planets Of primary importance are the planetary grav-
often show great similarity to Earth’s. In other ity, atmospheric pressure and density, with
cases, such as cryovolcanism on the icy moons ambient temperature a secondary effect. (For
of outer planets, the differences are striking: detailed reviews of how ambient conditions
instead of silicate magma, icy mixtures erupt in affect volcanism, see Zimbelman and Gregg,
extremely low-temperature environments. The 2000). In addition, magma compositional varia-
crusts of these moons are composed primarily tions among the planets can strongly influence
of water-ice, and may host liquid water bodies the expression of volcanism at the surface.
in the subsurface. Therefore, ices are analogous
to silicate rocks on Earth. On many of these 17.2.1â•… Gravity
worlds we see evidence for both active and past In modeling terrestrial volcanic processes, it is
cryovolcanic eruptions, which are at times unex- easy to forget that gravitational acceleration is
pectedly Earth-like in expression. in fact a variable that depends on the mass of
On the terrestrial planets (Mercury, Venus, the individual planet (Table 17.1). In practice,
Mars), as well as Earth’s Moon and Jupiter’s Io, this affects everything from the depth of melt-
where silicate volcanism has occurred or is cur- ing in a planet’s mantle to the height pyroclasts
rently occurring, the processes bringing material might reach in an explosive eruption. To illus-
to the surface are similar to terrestrial processes. trate, assuming that Earth’s mantle is composed
Therefore, models used to understand Earth’s of “wet” lherzolite, this material will begin to
volcanism are generally applicable to these other melt at a depth of ~75 km. On Mars, a similar
bodies, although model parameters are different. melting pressure is reached at a depth almost
Cryovolcanism requires significantly different three times greater (~225 km), because of the
models, in particular because of the difficulty low gravity. This has implications for the min-
of cryomagma ascent given the relative dens- imum ascent rate magma would require to
ities of water and ice. This chapter shows how reach the surface before solidifying, which in
models of volcanic processes, accounting for the turn sets limits for the minimum eruption rates
different ambient conditions and magma com- (Wilson and Head, 1981).
positions, can be used to understand volcanic Low gravity can also influence the dispersal
behavior of planetary bodies, and uses Mars, of pyroclastic ejecta. On the Moon, for example,
Io, and Enceladus as examples of the diversity dark mantle deposits (DMDs) are interpreted to
of volcanism in the Solar System. For consider- be pyroclasts emplaced during explosive erup-
ation of environmental factors influencing vol- tions (e.g., Wilson and Head, 1981; Weitz and
canism on Venus, see Grosfils et€al. (2000), Stofan Head, 1999). The most likely origin for these fea-
and Smrekar (2005), and also the discussion of tures is that they are pyroclasts generated from
volcanism in the high-pressure sea-floor envir- volatile-bearing magmas that would have gener-
onment in Chapter 12. Recent results about vol- ated hawaiian-style lava fountains (and associ-
canism on Mercury revealed by the MESSENGER ated scoria cones) if they had erupted on Earth.
spacecraft are discussed by Head et€al. (2011) and Instead, the lower lunar gravity, coupled with a
recent lunar results from the LRO spacecraft are lack of atmosphere, allowed the lava fragments
discussed by Joliff et€al. (2011). to be dispersed over distances that precluded
the building of a proximal volcanic cone.
It has long been theoretically argued that
17.2╇ ╇Parametric differences among low ambient gravity should result in thicker
planetary bodies lava flows, all other parameters being equal
(e.g., Fink and Griffiths, 1990). Thus, basaltic lava
Volcanic processes, from melt generation flows on the Moon should be thicker than those
and ascent to the eruption and deposition of on Mars; basaltic lava flows on Mars should be
9780521895439c17_p384-413.indd 386
Object radius (km) (m s−2) (kg m−3) Temperature (K) Pressure (Pa) products Expressions of volcanism
Earth 6371 9.81 5515 288 1.0 × 105 silicate Effusive and explosive
Moon 1737 1.6 3300 277 3 × 10 –10 silicate Mostly effusive, some
explosive
Mercury 2440 3.70 5427 100–725 5 × 10 –10 silicate Mostly effusive, some
explosive
Venus 6051 8.87 5204 733 9 × 106 silicate Mostly effusive
Mars 3389 3.73 393 215 600 silicate Mostly effusive, some
explosive
Io 1822 1.80 3528 85 (night) to < 10 –4, silicate, some sulfur Mostly effusive, some
140 (day) higher at possible explosive
(110 ave.) locations
of plumes
Europa 1565 1.31 3030 102 10 –6 H2O, unknown other Effusive cryovolcanism,
constituents possibly some explosive
Ganymede 2634 1.43 1940 110 < 10 –6 H2O, unknown other Effusive cryovolcanism
constituents
Enceladus 249 0.11 1120 75 trace, H2O, unknown other Explosive (plumes)
variable constituents, probably
ammonia
Titan 2575 1.35 1881 94 1.5 × 105 H2O, unknown other Possibly both effusive and
constituents, probably explosive cryovolcanism.
ammonia, methanol
Triton 1353 0.78 2054 38 1.4–1.9 water with unknown Effusive cryovolcanism,
other constituents, explosive cryovolcanism
probably nitrogen as a result of solar
green-house activity
8/2/2012 10:06:06 AM
P LA NETA RY VO LC A NISM 387
and bodies without insulating atmospheres can high thorium concentrations of some volcanic
experience hundreds of degrees in temperature areas on the Moon (e.g., Jolliff et€al., 2011), sug-
variation in the course of a single revolution. gest compositions that may approach rhyolitic.
Ambient temperature affects the rate of heat In this context, rhyolitic compositions suggested
transfer from eruptive products to their sur- for Venus (Fink et€ al., 1993; Treiman, 2007), to
roundings. Heat flux due to radiation varies as explain the physical properties of both tesserae
Thot4 − Ta4, while convective and conductive fluxes and pancake domes, may not be so far-fetched,
are both proportional to Thot€– Ta (Chapter 5). In and may have been facilitated by crustal recyc-
general, this implies that magmas erupted into ling due to catastrophic resurfacing of the crust,
cooler environments should cool more quickly, or a past wetter environment.
as might be expected; however, an exception There is much less known about the chemis-
to this rule is Venus, whose high atmospheric try of cryomagmas. For most icy bodies, cryomag-
density implies effective convective heat trans- mas are inferred to have low-viscosity, water-rich
fer (Section 17.2.2), despite the high ambient compositions. Given the lack of atmospheres in
temperature. these cases, explosive activity remains possible.
An exception may be Titan, where the possi-
17.2.4â•… Magma compositions bility of large abundances of liquidus-deflating
The lack of plate tectonics on all rocky worlds ammonia in the mantle may produce rheologies
other than Earth appears to have resulted in similar to those of terrestrial basalts or andes-
much less petrological diversity, as a result of ites (Kargel et€al., 1991). Although such proper-
less crustal recycling. For worlds where past or ties may result in larger volatile concentrations
present silicate volcanism has been detected, than on other icy satellites, the presence of a
mafic compositions (i.e., basalts) appear to dom- thick 1.5-bar atmosphere (0.15 MPa; Table 17.1)
inate. Relative to more felsic (silica-rich) mag- may act to suppress explosivity.
mas, mafic magmas have higher densities, and
lower viscosities and volatile contents. Taken
together with the apparent relative paucity of 17.3╇ ╇╇Volcanism on Mars
water, the primary volatile in the most explo-
sive volcanic eruptions on Earth, this suggests a 17.3.1â•…Observations of volcanic features
tendency towards the more effusive end of the on Mars
volcanic spectrum. There are, however, some Mars has been the main focus of planetary
exceptions. exploration since the 1990s and, therefore, the
High-temperature, low-viscosity volcanism data set available for studying its geologic proc-
on Io might be best explained by ultramafic esses is vast, ranging from orbital observations
compositions (Williams and Howell, 2007). to in situ measurements from landers and rov-
Geochemical analysis of Mars shows some evi- ers. The martian shield volcanoes are the lar-
dence for both ultramafic and more felsic mag- gest in the Solar System, and among the wide
mas, including basaltic-andesitic, andesitic variety of volcanic features are vast lava plains,
and possibly even dacitic surface chemistries many channels, shields, domes, and cones, and
(Section 17.3.2). It has been argued that weath- evidence of extensive pyroclastic deposits (Figs.
ering processes could produce high-silica sur- 17.2, 17.3; Greeley and Spudis, 1981; Greeley
face compositions on Mars, but it seems likely, et€ al., 2000; Carr, 2006). The presence of a fro-
given the large inferred sizes of martian magma zen cryosphere, 3–5 km thick, trapping water
chambers, that chemical differentiation of mag- in deep aquifers, is thought to have promoted
matic bodies in the crust would produce some magma–ice and magma–water interactions,
volume of intermediate- to high-silica magmas. and some surface features may have resulted
On Earth, the most felsic magmas (e.g., rhyo- from these interactions (Fig. 17.3(f); Fagents
lites) are the result of plate tectonic processes et€al., 2002; Wilson and Mouginis-Mark, 2003a,b;
and recycling of a water-rich crust. However, Wilson and Head, 2004).
The most prominent volcanoes on Mars are sheet lava flows, and deposits that have been
the giant shield volcanoes located in the Tharsis interpreted as pyroclastic (Mouginis-Mark et€al.,
volcanic province, a vast plateau ~8000 km in 1988). Interestingly, Alba Mons is antipodal to
diameter that covers ~25% of the planet’s sur- the Hellas impact basin, and early seismic wave
face (Fig. 17.2). Within this region, Olympus propagation modeling suggested that the energy
Mons, the tallest volcano in the Solar System, of the Hellas impact was sufficient to break the
is > 500 km wide and its summit reaches 25 km antipodal crust, creating fractures that could
above mean planetary radius, with maximum have acted as volcanic conduits that enabled the
flank slopes of 5˚. It is dominated by tube- and formation of Alba Mons at its present location
channel-fed lava flows of likely basaltic com- (Williams and Greeley, 1994).
position (Bleacher et€al., 2007), with a series of The second major volcanic area is the
nested summit calderas > 60 km across, and evi- Elysium Province (Fig. 17.2), which is dominated
dence of past glacial ice (Neukum et€ al., 2004). by Elysium Mons, Hecates Tholus, and Albor
Tharsis is also home to three smaller shield Tholus, and their surrounding lava flow fields.
volcanoes–Arsia, Pavonis, and Ascraeus Montes Hecates Tholus (Fig. 17.3(b)) is ~160 × 175 km
(Fig. 17.3(a)€ –and their associated rift aprons wide and extends ~6 km above mean planetary
and small shield fields (Fig. 17.3(e); Crumpler radius, and its flanks are dissected by shallow
and Aubele, 1978; Hauber et€ al., 2009). The radial valleys thought to be formed fluvially as
region also has vast interconnecting lava plains a result of melting of ice by subsurface magma
(Fig.17.3(d)), and several smaller (< 200 km diam- bodies (Fassett and Head, 2006). The dissected
eter) volcanoes named tholi (sing.: tholus, Latin nature of Hecates Tholus suggests it may be
for “cupola” or “dome”) which may be shields composed of ash or other easily eroded pyro-
whose lower slopes were buried by lava (Greeley clastic deposits (Mouginis-Mark et€ al., 1982), as
and Spudis, 1981). opposed to Elysium Mons, which appears com-
A unique volcanic structure in north Tharsis posed of lava flows.
is Alba Mons (Fig. 17.2), which covers an area The third major volcanic area on Mars is the
larger than Olympus Mons, but has flank slopes Circum-Hellas Volcanic Province, which is dom-
of less than 1˚ and thus lacks the relief of shield inated by Tyrrhenus and Hadriacus Montes, and
volcanoes. The flanks are heavily fractured, indi- Amphitrites Patera, and their associated flow
cating the influence of regional stress patterns. fields, surrounding the Hellas impact basin in
Alba Mons contains two discrete caldera-like the southern hemisphere (Fig. 17.2; Williams
features, numerous channelized, tube-fed and et€ al., 2009). The heavily channeled and
dissected nature of these shields (Fig. 17.3(c)) information derived from a variety of sources,
suggests that they are composed of easily ero- but there remain gaps in our understand-
dable, friable pyroclastic deposits rather than ing. Nonetheless, from a physical volcanology
lava flows (Greeley and Crown, 1990; Crown standpoint, we can make some fairly confident
and Greeley, 1993). The caldera-like depressions assessments of the composition of Mars’ surfi-
and surroundings of Peneus, Malea, and Pityusa cial deposits, and hence of its magmas.
Paterae suggest that they are similar to terres-
trial “supervolcanoes”, and produced ignim- Compositional data from spacecraft
brite deposits that have since been modified Compositional data in the form of spectra from
by fluvial, aeolian, and periglacial processes orbiting spacecraft have been collected by Mars
(Williams et€al., 2009). Global Surveyor (MGS), Mars Reconnaissance
Age estimates based on impact crater size– Orbiter (MRO), and Mars Express (ME). Spectral
frequency distributions suggest that most of the data are all consistent with a martian surface
volcanic activity in the Circum-Hellas Province composed primarily of mafic (probably bas-
is significantly older than elsewhere, between altic) materials that have been locally altered
~3.5 and 4.0 Ga, with little evidence that any (Bandfield et€al., 2000; Bandfield, 2002; Boynton
activity occurred more recently than 1 Ga. In et€al., 2004, 2007; Poulet et€al., 2007). Early spec-
contrast, effusive volcanism appears to have tral studies studies (Bandfield et€ al., 2000) sug-
dominated the later part of Mars’ volcanic his- gested a component of the martian surface that
tory: crater-count ages suggest that the Tharsis is consistent with an evolved magma compos-
shields are younger than ~1.8 Ga, whereas some ition€ – possibly andesitic. However, an equally
flows in the volcanic plains between Tharsis and valid interpretation is that the component is
Elysium may be as young as 10 Ma (Hartmann altered basaltic glass.
and Berman, 2000). Inside Nili Patera, the caldera on top of Syrtis
Major, Christensen et€ al. (2005) found dacitic
17.3.2â•…Composition of Mars volcanic rocks, with 60–63 wt.% SiO2. The outcrop is small
materials (a few square kilometers). Skok et€al. (2010) sug-
There are four main bodies of evidence for the gest that at least some of this enriched silica com-
composition of the martian surface: (1) remotely ponent may be caused by hydrothermal activity,
sensed spectral data from orbiting spacecraft; and is not an outcrop of primary igneous rocks.
(2) martian meteorites; (3) spectral data from Regardless, the total outcrop area is small enough
Martian landers and rovers; and (4) morphologic to make this region an interesting anomaly
data. Figure 17.4 summarizes compositional rather than representative of martian volcanic
Martian meteorites
Martian meteorites are also called the
“shergottite-nakhlite-Chassigny (SNC)” group,
referring to the names this class of meteorites
received before it was recognized that they orig-
inated on Mars. As of 2011, 97 meteorites have
been positively identified as originating on Mars.
Although there are other clues to a martian ori-
gin, the “smoking gun” is the composition of
gases trapped within solidified melt that is itself
trapped within the meteorite: the composition
of these trapped gases matches the composition
of the martian atmosphere as measured by the
Viking Landers (e.g., Treiman et€al., 2000).
The SNC meteorites are all mafic or ultra-
mafic igneous rocks. They range in compos-
ition from basalts (some shergottites) to dunite
(Chassigny). The nahklites are ultramafic wehr-
lites or clinopyroxenites; some shergottites are
lherzolites or harzburgites (McSween, 1994).
The basalt shergottites are interpreted to come
from surface lava flows, whereas the remaining
martian meteorites likely came from magmatic Figure 17.5╇╇ (a) Variation of pressure in the upper crust
intrusions (Nyquist et€ al., 2001), although the for Earth and Mars. Solid curves show the total pressure
felt at any depth (the sum of lithoststatic and atmospheric
precise locations of origin are as yet unknown
pressure); dashed curves show the lithostatic pressure only.
(Hamilton et€al., 2003).
Nucleation (N) and fragmentation (F) depths are indicated
by arrows for a basaltic magma initially containing 0.3 wt%
Lander and rover data dissolved H2O. (b) Crustal density as a function of depth for
To date, data on the composition of in situ Mars (solid curves) and Earth (dashed curves). Curves are
martian samples have been returned from marked with the porosity (%) of surface rocks.
the Viking Landers 1 and 2, Mars Pathfinder’s
Sojourner, the Mars Exploration Rovers Spirit
and Opportunity, and Phoenix. The Viking affected by sample contamination with weath-
landers’ X-ray fluorescence results for Mars ering products, given that Sojourner had no
soil revealed a composition similar to that of means to remove the weathering rinds.
iron-rich basalt. Sojourner’s alpha-proton X-ray Spirit and Opportunity were each equipped
spectrometer (APXS) collected compositional with a Rock Abrasion Tool (RAT), designed to
data from the weathered surfaces of boulders remove surficial weathering products prior to
at the landing site. McSween et€ al. (1999) used examining the fresh rock face beneath. The data
these data to calculate the composition of a collected have revealed a wealth of geologic proc-
likely end member martian igneous rock, and esses affecting the surface (e.g., Ruff et€al., 2006;
determined that such a rock might be an andes- Squyres et€al., 2006). However, the measured soil
ite. They stress, however, that these results and rock compositions remain most consistent
are model-dependent, and could be strongly with a mafic or ultramafic igneous parent.
The Phoenix lander was designed to exam- dacitic; Wichura et€al., 2010), drained lava tubes
ine the composition of martian ice rather than as observed on Mars have only been seen in ter-
martian rock (Arvidson et€ al., 2009). However, restrial basalt flows. Thus, the presence of mar-
the mechanical behavior of the soils and their tian lava tubes and channels supports a mafic
appearance are consistent with those observed lava composition.
at the Viking 2 (Arvidson et€ al., 2009). These In contrast, however, multiple studies have
results are consistent with the soils being a examined the morphologies of specific martian
weathering product from a mafic or ultramafic lava flows and inferred a more evolved compos-
igneous parent rock. ition, including rhyolite (e.g., Hulme, 1976; Fink,
1980; Zimbelman, 1985; Wadge and Lopes, 1991;
Morphology Warner and Gregg, 2003). It is important to note,
Of all the available data sets, morphology is however, that the physical characteristics of
the least reliable for determining lava compos- lava flows (typically flow thickness and surface
ition. Observations and laboratory simulations textures) that have led to inferred evolved lava
of terrestrial volcanic behaviors indicate that compositions can be created by emplacement
lava rheology exerts a strong control on vol- kinematics and rheology. In other words, infer-
canic behavior and morphology (see Chapter 12, ring lava flow composition from flow morph-
Fig. 12.6, for example). Early experiments sug- ology is difficult: rheology and pre-existing flow
gested that SiO2 content can be directly linked surfaces exert strong controls on lava flow morph-
with lava rheology (Hulme, 1974, 1976); we now ology, and rheology is not uniquely related to lava
know that this is an oversimplified interpret- composition. Added to these considerations are
ation. Nonetheless, volcanic morphologies that the complicating effects of Mars’ environmental
are apparently generated by low-viscosity flows factors (Sections 17.2, 17.3.4).
are most easily explained by lavas with a mafic The shield morphologies of many of the
or ultramafic composition. martian volcanoes are similar to those of ter-
Lava tubes and lava channels, for example, restrial shield volcanoes (e.g., Mauna Loa,
have been observed on Mars (Bleacher et€ al., Hawaii) and all terrestrial shield volcanoes
2007), as have thin, fluid flood-style flows are basaltic. Furthermore, the morphologies
(Jaeger et€ al., 2007). Although channels are of smaller volcanic features and lava flows
observed in lava flows with a range of compo- are generally suggestive of low-viscosity mag-
sitions on Earth (e.g., ultramafic, phonolitic, mas. Even the putative pyroclastic deposits
surrounding Tyrrhena and Hadriaca Montes
could be generated by mafic magmas (Greeley
Figure 17.6 (cont.)╇ ╇ km across) is partly covered with and Crown, 1990; Crown and Greeley, 1993),
dark material, presumably lava, that appears significantly with explosive activity driven by either mag-
warmer in Near Infrared Mapping Spectrometer data than matic volatiles or water/ice encountered by
the adjacent orange-colored “island,” parts of which are cool rising magma erupting into a low-pressure
enough for SO2 to condense. Image PIA03601, NASA/JPL. atmosphere.
(e) The source of the Prometheus plume lies at the distal Thus, the preponderance of evidence is that
(westernmost) end of a compound pahoehoe flow field. Mars has generated ultramafic to mafic volcanic
Image PIA02565, NASA/JPL/University of Arizona. Inset box
products throughout its history. More evolved
shows the location of (f), which is a high-resolution view of
volcanic materials (such as the high-silica mate-
the active flow front. Dark areas on the flow show recent
lava breakouts, older flow surfaces brghten with time due to rials found within Nili Patera) are volumetric-
deposition of SO2 front from the plume. Fresh SO2 frost is ally insignificant on the Martian surface.
also seen on the surrounding terrain. Image PIA02557, NASA/
JPL/University of Arizona. (g) Bright lobate features to the 17.3.3â•… Magma ascent processes on Mars
east of Emakong Patera have been interpreted as fresh sulfur The low pressure of the martian atmosphere
flows. Image PIA02539, NASA/JPL/University of Arizona. See (Pa = 600 Pa; Table 17.1), together with the low
color plates section. acceleration due to gravity (g = 3.73 m s−2), exert
strong controls over magma ascent and erup- that the pressure at any depth is less on Mars
tion. At depth, warm, buoyantly ascending than on Earth (Fig. 17.5(a)). This implies that the
diapirs may undergo pressure-release melting rate of densification of country rock with depth
to produce partially molten magma reservoirs. is also lower, i.e., crustal rocks at a given depth
The buoyancy force Fb on a given volume V of on Mars will be less compacted than on Earth
melt is (Fig.17.5(b)). Therefore, an ascending magma will
encounter country rock with a density equal to
Fb = g(ρr–ρm)V, (17.1)
its own at greater depths on Mars than on Earth,
where ρr and ρm are the country rock and magma thus implying deeper neutral buoyancy zones
densities, respectively. Since magma ascent rate by a factor of ~4 (Wilson and Head, 1994).
depends on the buoyancy force (which depends Magma chambers forming at both deep
upon gravity), diapirs of a given volume would rheological barriers and shallow neutral buoy-
ascend more slowly on Mars than on Earth. ancy zones will both be correspondingly deeper
Conversely, melt bodies would have to be larger on Mars than on Earth. This has a number of
to avoid excessive cooling and stalling during implications. First, higher driving pressures and
buoyant ascent, implying larger volumes would wider dikes are needed for an ascending magma
be available for eruption (Wilson and Head, to reach to the surface from greater depths. The
1994) on Mars than on Earth. high pressure gradients driving the magma lead
Ascending melt may stall at a rheological to higher velocities, larger mass fluxes, and cor-
barrier (e.g., the base of the lithosphere), or at a respondingly larger volume eruptions, despite
density barrier, where the country rock density the lower buoyancy forces on Mars. In addition,
becomes equal to that of the melt (i.e., the level consideration of fracture mechanics suggests
of neutral buoyancy). In either case, a magma res- that the widths w and horizontal extents H
ervoir may form. The depth to the base of the of dikes are inversely proportional to gravity
lithosphere is governed by the cooling history raised to some negative power that depends
of the planet, which is related to the size of on the model used, e.g., w ∝ g−1/3 and H€€∝ g−2/3
the body. Given that Mars’ diameter is roughly (Wilson and Head, 1994; see also Chapter 3).
half that of the Earth, its lithosphere is likely This implies that dikes on Mars would be sys-
to have thickened more rapidly and produced tematically wider and longer by factors of ~1.4
a deep rheological barrier to magma ascent. If and 1.9, respectively, than dikes on Earth. For
the magma is still buoyant on encountering laminar flow, ascent velocity is proportional to
and stalling at such a boundary, and the stress square of the dike width w and the driving pres-
regime (due to buoyancy pressure, reservoir sure gradient, and the mass (or volume) flux is
excess pressure, and external stresses) allows proportional the product of the velocity and
chamber walls to rupture, magma may continue cross-sectional area of the dike (wH). This leads
to ascend through propagating dikes. When a to a strong inverse relationship (~g−5/3) between
level of neutral buoyancy is reached at shallower eruption rate and gravity, indicating a system-
levels in the crust, in the absence of excess driv- atic trend to higher eruption rates by a factor
ing pressure, magma may stall once again. of ~5 on Mars compared to Earth (Wilson and
The total pressure P felt at any depth beneath Head, 1994). Taken together, the influence of
the planetary surface is the sum of the lithos- greater dike dimensions and deeper magma
tatic pressure, ρrgz, and the external atmos- chambers may account for a factor of seven
pheric pressure, Pa. At depths greater than ~10 times greater mass fluxes for eruptions on Mars
m on Earth, atmospheric pressure is a negligible than on Earth, all other factors being equal.
component of the total pressure. On Mars, the This clearly has implications for the volumes
external pressure is so low as to be insignificant of material erupted, and for the lengths of lava
(Fig. 17.5(a)). Of greater importance at depth is flows and the sizes of the edifices constructed.
the low martian gravity, which produces lower As magma reaches shallower levels in the
pressure gradients dP/dz in the lithosphere, such lithosphere, the pressure drops to the point
at which magmatic gases (H2O, CO2, etc.) start rapidly decompress down to the low ambient
to exsolve. For a given magma composition, pressure, leading to a much greater energy
exsolution will tend to occur at depths ~2.5–3 release per unit mass of gas than would be the
times greater on Mars than on Earth (Wilson case on Earth, and accelerating the mixture to
and Head, 1994), primarily because of the lower significantly higher velocities. Wilson and Head
lithostatic pressure gradient. However, on Mars, (1983) give a simplified expression showing
the low lithostatic pressure environment is that the square of the eruption velocity at the
likely to lead to fragmentation of magma con- vent is proportional to the natural logarithm
taining relatively modest amounts of volatiles. of the ratio of the pressure at fragmentation to
As discussed in Chapters 4 and 7, fragmentation atmospheric pressure. This leads to velocities
of low-viscosity magmas may take place due to in sustained eruptions being a factor of 1.6–2
(1) rapid bubble expansion leading to hydro- times greater on Mars than on Earth for similar
dynamic breakup, (2) expansion exceeding the ascent conditions.
structural relaxation rate of the melt, leading
to brittle fragmentation, or (3) exceeding a crit- 17.3.4â•… Eruption styles and processes
ical bubble volume fraction (typically taken as Our understanding of the physics of magma
65–85 vol. %). If a critical bubble volume fraction ascent suggests that in general one might expect
is the dominant criterion for fragmentation, the a greater propensity for explosive volcanism on
depth at which this threshold is reached (for a Mars, given magmas with even modest volatile
given volatile content and subsonic conditions), contents. The potential for ascending magma to
is a factor of ~3 deeper on Mars because the crit- interact with subsurface ice only acts to increase
ical fragmentation pressure is found at greater the chances for explosive volcanism.
depths (Fig. 17.1; Wilson and Head, 1994). The Assuming that some Mars magmas had suf-
fragmentation depth is non-trivial to determine ficient volatiles for fragmentation to occur, we
if the eruption is supersonic (Mitchell, 2005), consider the manifestation of explosive volcan-
but will still be greater in depth than on Earth. ism in the martian environment. Vigorous frag-
The rapid ascent and expansion rates expected mentation and high-velocity eruption of finely
in the low-pressure martian environment mean fragmented tephra may have preferentially
that the brittle and hydrodynamic breakup produced plinian eruption columns rather than
criteria would also be met at correspondingly the hawaiian lava fountains typical of basaltic
greater depths on Mars. Even if martian mag- explosive activity on Earth. Early plume models
mas contained fewer volatiles than typical indicated that, provided the erupting mixture
terrestrial mafic magmas, calculations using attained buoyancy above the vent, convective
simple solubility laws suggest that CO2 or H2O columns could rise a factor of five times higher
contents exceeding ~0.01 wt% may lead to frag- on Mars than similar mass flux eruptions on
mentation on Mars (cf. ~0.07 wt% on Earth) dur- Earth, implying rise heights of tens to >100
ing sustained magma ascent in which bubbles km, which would serve to distribute material
and magma are dynamically coupled (Wilson widely around the planet (Wilson and Head,
and Head, 1983). 1994). However, the low density of the martian
The low lithostatic pressure gradient means atmosphere would cause difficulty in achiev-
that nucleation and fragmentation levels are ing buoyancy because of the limited potential
both deeper and more widely separated on for heating and expansion within the column,
Mars than on Earth (Fig. 17.5(a); Wilson and such that collapsing columns and pyroclas-
Head, 1994). This leads to an extended period tic density currents (PDCs) would be common
of bubble growth, the potential for numerous for a wider range of conditions than on Earth.
nucleation events and bubble populations, and A later reconsideration of assumptions under-
ultimately a more finely fragmented magma. lying Morton-style convective plume models
Once the bubbly magma has transitioned to (see Chapter 8), found that some of the basic
a particle-laden gas stream, the mixture will model assumptions are readily violated under
present martian atmospheric conditions, such In the case of lava effusion, eruption rates
that column heights would be limited to ~10–12 may have been considerably greater than is typ-
km because the low-density atmosphere can- ical on Earth. Indeed, large shields and exten-
not support convection to significant altitudes sive lava plains are common on Mars (Section
(Glaze and Baloga, 2002). While PDCs could cer- 17.3.1; Fig. 17.3). Once erupted on the surface
tainly have been generated from failed convect- of Mars, lavas would experience a combination
ive columns, one might also envision a style of of radiative and convective cooling from the
particle fallout more akin to that observed on flow surface at temperature Thot, and conductive
Io (Section 17.4), starting from heights above cooling to the substrate below (Section 17.2.3;
the limits of convective capability, and produ- Chapter 5). Given the low atmospheric density,
cing diffuse deposits (Wilson and Head, 2007). A the radiative heat flux would dominate over con-
putative, past dense martian atmosphere would vective cooling. Furthermore, the total heat flux
have permitted convective ascent to greater col- at a given surface temperature would be less
umn heights, and more widespread deposition. than on Earth. This, together with the greater
It is conceivable that the ancient martian vol- predicted thicknesses of martian flows (a com-
canoes interpreted to be composed of explosive bination of high effusion rates and low gravity),
deposits (Section 17.3.1) were produced by vig- means that flows would in general cool more
orous eruptions during periods of high atmos- slowly and travel farther than terrestrial flows.
pheric density. Taken together, greater predicted dike widths
Hawaiian-style lava fountains (Chapter 8), (× 2) and effusion rates (× 7) for eruptions in the
if they formed in less vigorous eruptions, are martian environment, imply flows five times
likely to have contained finer grain sizes, been longer than on Earth for similar magma compo-
erupted at greater speeds, and dispersed mater- sitions (Wilson and Head, 1994).
ial over greater areas than for equivalent magma The discussion above deals largely in gen-
mass fluxes and volatile contents on Earth. This eralizations based on the expected influence
is likely to have led more commonly to broad of the martian gravitational and atmospheric
tephra cones rather than steep, narrow spatter environment; in reality there are certain to have
ramparts or coalescence of pyroclasts into lava been many variations on the styles of volcan-
flows (Wilson and Head, 1994). However, rela- ism predicted based on theoretical treatments.
tively few primary tephra cones have been iden- However, the general picture gleaned from
tified. One might infer that in general Mars’ lava observations of martian volcanic features is of
flow fields were formed by passive lava effusion, older explosive volcanism in the south transi-
or only very weaklyfragmented, volatile-poor tioning to younger effusive volcanism in the
magmas that rapidly coalesced to produce north; this is hard to reconcile with our under-
flows. standing of the physics of ascent and eruption
Strombolian activity (Chapter 6) may not in the martian environment. Although a denser
have been very common on Mars, based on the- early Mars atmosphere might have promoted
oretical considerations. Relatively low buoyancy widespread ash dispersal from convecting col-
forces on ascending bubbles, together with the umns, it would also have acted to suppress ves-
likely more rapid magma ascent speeds, would iculation and fragmentation to a greater degree
have acted to inhibit the decoupling of the than the latter-day tenuous atmosphere. One
bubbles from the melt that is required for slug possible explanation for the transition in erup-
formation and strombolian explosions. In the tive style might lie in a general depletion of
case where volatiles do decouple from a stalled magmatic volatiles over the period that Mars
magma column, the great expansion potential was volcanically active; early volatile-rich explo-
of the released gases in the martian environ- sive eruptions may have given way to volatile-
ment would produce wider and lower cones poor effusions. Mars’ small size and the lack of
than would be observed for terrestrial strombol- recycling of volatiles via plate tectonics may
ian activity. both have promoted irreversible gas loss from
the interior of the planet. Another possibility of which were interpreted to be a result of
is that patterns of explosive vs. effusive volcan- shield-building basaltic volcanism (Schaber,
ism are related to the availability of external 1980). The most common type of volcanic fea-
volatiles such as ground ice or groundwater, ture on Io is the patera (Fig 17.6(a), (d), (e), (g)).
and their propensity to promote explosive Although the origin of Ionian paterae is still
magma–water interactions. Spatial or tem- somewhat uncertain, they are thought to be
poral variations in the distribution of crustal similar to terrestrial volcanic calderas, formed
H2O reservoirs may therefore have had a strong by collapse over shallow magma chambers
influence on the manifestation of volcanism in following partial removal of magma. Some
the martian environment. Understanding this paterae show angular shapes that suggest
enigma remains a key objective for researchers structural control, indicating that they may be
in Mars volcanology. structural depressions that were later used by
magma to travel to the surface. More than 400
Ionian paterae have been mapped (Radebaugh
17.4╇ ╇╇Volcanism on Io et€ al., 2001; Williams et€ al., 2011). Their aver-
age diameter is ~40 km but Loki, the largest
17.4.1â•…Observations of volcanic features patera known in the Solar System, is > 200 km
on Io in diameter. The larger sizes of the Ionian fea-
Io’s colorful surface is studded with volcanic tures probably reflect large, relatively shallow
features and deposits (Fig. 17.6), and Io is so far magma chambers (Leone and Wilson, 2001).
the only body outside Earth known to exhibit Io’s surface shows some remarkably large
active silicate volcanism. Active eruptions were lava flow fields; the Amirani lava flow field, at
discovered in 1979 when the Voyager spacecraft ~300 km long, is the largest active flow field
revealed enormous, umbrella-shaped plumes known in the Solar System. Repeated imaging
(Morabito et€ al., 1979) and thermal anomal- of Amirani during the Galileo fly-bys allowed
ies (Pearl et€ al., 1979). Results from the Galileo eruption rates to be estimated at 50–500 m3 s−1
mission, which observed Io from 1995–2001, (Keszthelyi et€ al., 2001). The Prometheus flow
substantially advanced our understanding of field (Fig. 17.6(e)) extended 80 km to the west
volcanism on Io (Lopes and Williams, 2005). between Voyager and Galileo observations; cal-
Distant observations of Io were also made by the culated resurfacing rates are an order of magni-
Cassini spacecraft on its way to Saturn during tude lower than Amirani’s flow field. This range
2000–01 (Radebaugh et€al., 2004) and by the New of effusion rates is surprisingly modest given
Horizons spacecraft in 2007 (Spencer et€al., 2007) the very large size of these flow fields.
en route to Pluto. Ground-based monitoring has Io’s plumes exhibit significant variabil-
also been important in the study of Io’s volcan- ity. They range from < 100 km to > 450
ism (e.g., Marchis et€al., 2000; Veeder et€al., 1994). km in height. Large plumes (e.g., Pele and
More than 150 active volcanic centers have been Tvashtar; Fig. 17.6(c)) are associated with high-
identified from these combined observations; �temperature thermal anomalies, may be quite
there are likely to be many more that have faint (implying a vapor-rich composition), and
not yet been observed erupting. Recent results may produce annular deposits > 1200 km in
using data from the magnetometer on Galileo diameter, that are reddish or sometimes black
(Khurana et€al., 2011), suggest the presence of an in color (Fig. 17.6(b)). Smaller plumes, e.g.,
asthenosphere (global magma ocean) at least 50 Prometheus (Fig. 17.6(c)), tend to be associated
km thick under a low-density outer crust ~30– with relatively cool thermal anomalies, and are
50 km thick. optically more dense than Pele-type plumes,
Io’s volcanoes rarely build significant topo- implying a mixture of condensing gases and
graphic structures. There are only a few struc- particles. Deposits are typically pale yellow in
tures, called tholi, scattered across Io, some color. Intriguingly, during the 17-year interval
between Voyager and Galileo observations, consensus is that eruptions with temperatures
Prometheus’ plume moved ~80 km to the west characteristic of basalts (1400–1550 K) are com-
(see Section 17.4.4). mon on Io, but hotter, ultramafic volcanism may
be also present (Williams and Howell, 2007).
17.4.2â•…Composition of Ionian volcanic
materials 17.4.3â•… Magma ascent processes on Io
The main difference between the compositions Initial interpretations combining optical (Solid
of volcanic materials on Io and Earth is that Io State Imager, SSI) and infrared (Near-Infrared
contains negligible amounts of water (Carlson Mapping Spectrometer, NIMS) data of Io’s erup-
et€ al., 1997), and sulfur and sulfur dioxide are tions showed some eruption temperatures
the main volatiles detected in both plume mate- greater than those typical of terrestrial basalts
rials (McGrath et€al., 2000; Spencer et€al., 2000; (> 1450 K), with > 1850 K for at least one eruption
Jessup et€ al., 2004) and on the surface (Carlson (McEwen et€al., 1998; Davies et€al., 2001), which
et€al., 1997, 2007). Whereas Voyager observations exceeds any volcanic activity observed on Earth.
were unable to resolve whether sulfur or silicate Such temperatures would be consistent with an
was the predominant magma type, temperature undifferentiated and almost completely molten
measurements from the Galileo and the New interior, but this does not appear to be the case.
Horizons spacecraft clearly showed that most Models that couple tidal heating to volcanic heat
active volcanoes have temperatures too high to loss suggest that the interior should be no more
be erupting sulfur. A few locations show young, than ~20% molten (e.g., Moore, 2001), which
pale yellow or white flows (e.g., Emakong Patera; in turn suggests magmatic temperatures of
Fig. 17.6(g)) that may have been molten sulfur < 1550 K, regardless of geochemistry. The discrep-
(Williams et€ al., 2001, 2004), whereas others ancy between observation and models requires
may be silicates coated with sulfur. Greeley et€al. explanation. Three factors may be relevant: (1)
(1984) suggested that rising silicate magma may uncertainties in the thermo-spectral modeling
melt near-surface sulfur-rich country rock, pro- of eruption temperatures; (2) problems with geo-
ducing “secondary” sulfur flows, as opposed to physical models of interior thermal evolution;
“primary” flows that originate from molten sul- and (3) unusual conduit conditions.
fur magmas at depth. Reanalysis of SSI and NIMS data of two of
The main question regarding the compos- the more anomalous eruptive sites on Io, Pele
ition of Io’s silicate lavas is whether they are and Tvashtar, yield lower temperatures than
mafic (possibly superheated) or ultramafic. previously reported and account for 200–300
With no spectroscopic measurements of lava K of the 400 K gap between observation and
composition (mostly due to spatial resolution theory (Keszthelyi et€ al., 2007). The rest of
limitations), temperatures detected at active the difference can be explained by a quirk of
volcanoes provide the best clues to magma magma ascent in Io’s highly compressive litho-
composition. Temperatures of Io’s active lavas sphere. The possibility of superheating due
have been calculated from observations in the to viscous dissipation within the magma dur-
near-infrared from Galileo (e.g., Lopes et€ al., ing conduit ascent was raised by McEwen et€al.
2001; Davies et€ al., 2001; Howell and Lopes, (1998). This process is generally negligible for
2011), Cassini (Radebaugh et€al., 2004), and New basaltic magmas on Earth (e.g., Mastin, 1995),
Horizons data (Spencer et€ al., 2007). However, but has been suggested as potentially import-
temperatures determined from remote sensing ant for explosive silicic eruptions (e.g., Mastin,
data depend on the spatial resolution and wave- 2005). The unique stress distribution within the
length range; because lava cools rapidly after Ionian lithosphere (Jaeger et€al., 2003) requires
exposure at the surface, measured tempera- an ascending magma to overcome a confining
tures are unlikely to represent magmatic tem- pressure of ~0.5 GPa (Keszthelyi et€ al., 2007).
peratures (e.g., Lopes et€ al., 2001). The current For a magma ascending adiabatically against
friction and gravity, g, through a pipe of radius effusive, and those that are confined within
r, the change in temperature T with decreasing paterae. Plumes may be associated with all
depth z is given by: three eruption styles.
Explosion-dominated eruptions, typified by
dT n dP u 2 f
= + Pillan volcano’s 1997 eruption, are high effu-
dz ρ g c dz 2rc (17.2)
sion rate, vigorous outbursts of material dur-
ing short-lived (weeks-long) events. These
where P is the pressure, c is the bulk specific
events typically produce large (> 200 km high)
heat capacity, n is the mass fraction of vola-
plumes (known as Pele-type plumes; McEwen and
tiles driving the ascent, ρg is the density of the
Soderblom, 1983) likely due to release of juven-
volatile phase, u is the velocity of ascent, and
ile volatiles, and result in extensive (up to ~1200
f is the wall friction factor. The first term on
km diameter) red annular deposits of short-chain
the right-hand side represents adiabatic cool-
sulfur, and sometimes dark deposits, presum-
ing of any exsolved gas phases, and the second
ably silicate pyroclastics (Fig. 17.6(b)). In addition
term represents viscous dissipation. Assuming
to plumes, lava fountains may also form part of
turbulent flow (Wilson and Head, 2001) of
the eruption sequence (e.g., at Tvashtar in 1999;
an unvesiculated magma allows significant
Keszthelyi et€ al., 2001). Many of the dynamical
simplification:
features displayed by the plumes are expected
dT 1 dP g from the flow of gas out of a nozzle and into
~− − . a near-vacuum (Geissler and Goldstein, 2007).
dz ρ g c dz c (17.3)
Expansion of erupted gases in the low-pressure,
low-gravity environment readily explains the
which is independent of velocity. The g/c term
very high plumes observed. Given the absence of
is second order, and can be ignored here. Thus,
an atmosphere, the erupting mixture is unable
assuming 0.5 GPa overpressure (leading to ~0.34
to form a convecting column, and instead erupt-
GPa driving pressure) at 20 km depth, taking ρg
ing particles follow quasi-ballistic trajectories,
= 2700 kg m−3, g = 1.7 m s−2 and c = 840 J kg−1 K−1
influenced by particle–particle collisions and
gives a temperature increase of ~110 K.
shocks in the emerging gas phase. Application
It is important to note that this simplistic
of simple ballistics equations relating plume
treatment is appropriate only for unvesiculated
height to eruption velocity (h = u02/2g), yields
magmas. Exsolution and expansion of volatiles
eruption velocities > 1 km s−1 in some cases.
during ascent will inevitably produce a cooling
Theoretical treatment of plume thermodynam-
effect that will act to counteract the viscous
ics (Kieffer, 1982) and numerical modeling of
superheating under a wide range of circum-
plumes and their entrained particulates (Zhang
stances. Hence, this balance between heating
et€al., 2003) explain many of the finer details of
and cooling effects may result in considerable
plume characteristics.
variation in eruptive temperatures, generally
Flow-dominated eruptions, typified by
ranging between hotter effusive and cooler
Prometheus (Fig. 17.6(e)) and Amirani volca-
explosive eruptions.
noes, produce extensive lava flow fields. The
ability of lava to travel large distances at mod-
17.4.4â•…Eruption styles and processes erate effusion rates (< 5–500 m3 s−1), together
on Io with thermal profiles along the Prometheus
Io’s continuous activity has allowed general and Amirani flows (Lopes et€ al., 2001, 2004)
classification of eruption styles (Williams and and high spatial resolution images (Fig. 17.6(f);
Howell, 2007), studies of plume dynamics Keszthelyi et€al., 2001), suggest that these large
(Geissler and Goldstein, 2007), and determin- Ionian flow fields are emplaced as insulated
ation of lava effusion rates (Keszthelyi et€ al., (tube-fed) flows of low-viscosity lava, similar to
2001). Three main styles of eruption have terrestrial inflated pahoehoe flow fields (Hon
been identified, primarily explosive, primarily et€al., 1994). Flow-dominated eruptions are also
Saturn’s small moon Enceladus (mean radius interpretation has been disputed (Soderblom
252 km) presents the most dramatic evidence of et€ al., 2009). Moore and Pappalardo (2011)
cryovolcanic activity observed to date (Fig. 17.8). argue against cryovolcanism on Titan, suggest-
Plumes of water vapor and icy particles are ing that the lobate features can be explained
observed emanating from the highly tectonized as sedimentary deposits produced by fluvial
south polar region (Fig. 17.8(b); Porco et€ al., activity. A region called Sotra Facula may host
2006; Spahn et€ al., 2006; Spencer et€ al., 2006; the best evidence to date for cryovolcanism
Waite et€al., 2006). This region hosts four linear, (Fig. 17.7g; Lopes et€al., 2010b; Kirk et€al., 2010)
sub-parallel lineaments (known as sulci, or more in the form of a pair of 1-km-high mountains,
informally as “tiger stripes”) in the ice shell. The possibly cryovolcanoes, with adjacent flow-like
sulci are typically 500 m deep, 2 km wide and features and a pit > 1 km deep that could be a
flanked by ridges ~100 m high (Spencer et€ al., caldera or vent structure. However, the exist-
2009), and they exhibit higher surface tempera- ence of cryovolcanism on Titan remains under
tures (up to 180 K) than the surrounding regions debate€– the “smoking gun” for active cryovol-
(70–80 K; Spencer et€al., 2006). Plume vent loca- canism, in the form of an enhanced thermal
tions identified in high resolution (~7 m/pixel) signature or active plumes, has not yet been
images show that these areas are mantled by detected.
smooth deposits that extend along the length of Voyager observations revealed eruptive
the sulci, and which thin laterally (Fig. 17.8(f)). plumes as tall as 8 km on Triton (Smith et€ al.,
Vent locations correlate well with the warm- 1989; Soderblom et€al., 1990) but these may not
est locations along the sulci, and at least some qualify as volcanic. Smith et€al. (1989) suggested
of them persisted over a two-year observation that the gas venting may be driven by solar
period (Spitale and Porco, 2007). Cassini obser- heating and the subsequent vaporization of sub-
vations have also made it clear that Enceladus’ surface nitrogen. Other proposed mechanisms
plume activity forms Saturn’s faint E-ring (Fig. for gas venting include melting of or convection
17.8(a)). Away from Enceladus’ south polar within the solid nitrogen polar caps (Brown and
region, there is a variety of terrain types (Fig. Kirk, 1994; Duxbury and Brown, 1997), but these
17.8(c)), including smooth, resurfaced terrain. do not explain all of the observed plume fea-
However, no obvious cryovolcanic landforms tures. However, Triton also shows other evidence
have been observed. of cryovolcanic resurfacing, including smooth
Before the Cassini-Huygens mission, geo- plains, quasi-circular scarp-bounded depressions
chemical and geophysical models suggested similar to paterae, and various pits, domes, and
that Titan, Saturn’s largest moon (mean radius channels (Fig. 17.7(h); Croft et€ al., 1990). The
of 2575 km) may be cryovolcanically active, resurfaced terrains could be the manifestation
facilitated by substantial quantities of ammonia of high heat production in the interior. Triton’s
in the interior, which would help to maintain a inclined, retrograde orbit around Neptune sug-
subsurface liquid layer. Detection of 40Ar in the gests that it is a captured satellite (McKinnon,
atmosphere (Niemann et€al., 2005) supports the 1984). Therefore, as a result of stresses exerted
case for cryovolcanism, as it implies outgassing during orbital circularization, Triton may have
from Titan’s interior. Subsequently, Cassini has undergone significant tidal heating and interior
observed a range of features, including lobate, melting to produce the diversity of geologic fea-
flow-like morphologies, which have been tures observed today.
interpreted as evidence of cryovolcanic activ- Other icy bodies, including Uranus’ satellites
ity (Sotin et€al., 2005; Barnes et€al., 2006; Lopes Ariel, Miranda and Titania, also show intriguing
et€al., 2007; Wall et€al., 2009). Furthermore, the features that may have cryovolcanic origins,
Visible and Infrared Mapping Spectrometer perhaps facilitated by the presence of exotic
(VIMS) observed periodic brightening at two cryomagma compositions. The variations in the
locations that was attributed to active cryovol- geologic and eruptive activities among icy satel-
canism (Nelson et€al., 2009(a),(b), although this lites can be attributed to differences in thermal-
orbital evolution, internal composition, degree temperatures near the sulci. Amorphous ice
of differentiation, and ascent processes, which is found between sulci (Brown et€ al., 2006;
are presently poorly understood and may vary Newmann et€al., 2008), which is likely formed
from world to world. Additional data from by condensation of plume gases at the lower
future missions are required to fully decipher (⛪ 100 K) temperatures away from the active
the cryovolcanic histories of icy bodies. The eruptions (Spencer et€al., 2006).
remainder of this section focuses primarily on Direct sampling of Enceladus’ plumes by
Enceladus, because of its spectacular and well- the Cassini spacecraft shows that they consist
observed activity. predominantly (90–94 wt.%) of water vapor and
water-ice grains (Waite et€al., 2006; Hansen et€al.,
17.5.2â•… Compositions of cryomagmas 2011). Results from the Cosmic Dust Analyzer
The compositions of cryomagmas depend on showed that particles closer to the surface tend
the materials that condensed from the solar to be large and rich in sodium salts, whereas
nebula at a given distance from the Sun. In further away from the surface the particles are
addition to water, at Saturn’s orbit, methane small and salt-poor. Cassini’s Ion Neutral Mass
and ammonia condensed, so it is likely that Spectrometer (INMS) and Ultraviolet Imaging
these exist in the moons’ cryomagmas. In the Spectrometer (UVIS) obtained complemen-
further reaches of the Solar System, more vola- tary information on plume composition. Apart
tile compounds, such as carbon monoxide, car- from water, INMS detected carbon dioxide (5%),
bon dioxide, and nitrogen, may play a greater methane (0.9 %) and ammonia (0.8%), as well as
role. The viscosity of a cryomagma influences trace amounts of organics, H2S and 40Ar (Waite
the resulting morphologies of landforms it pro- et€al., 2006, 2009), and possibly molecular nitro-
duces. Liquid water would simply flood a sur- gen or carbon monoxide. UVIS results failed to
face, filling in depressions, but the inclusion of detect molecular nitrogen, however (Hansen
some amount of ammonia to form ammonia et€al., 2011).
hydrates would produce viscosities approach-
ing those of silicate lavas (e.g., Kargel et€ al., 17.5.3╅Cryomagma ascent and eruption
1991). This suggests that a broad spectrum processes on Enceladus
of eruption styles may be possible, and that Despite direct observations of active erup-
familiar landforms, such as lobate flows, shield tions in the outer Solar System (Enceladus and
volcanoes, domes or cones, might grace these Triton), the processes by which cryomagmas
distant landscapes. ascend on icy satellites are unclear. Many ascent
To date, the only direct data available on models have been based on the terrestrial mag-
the composition of cryovolcanic products are matic paradigm, and propose that cryomagmas
from Enceladus’ plumes. Near-infrared spec- rise buoyantly from the mantle into the lower
troscopy shows that Enceladus’ surface is ice shell, potentially stalling close to the surface
dominated by water ice, mainly in a crystal- to form a cryomagma chamber, and then erupt-
line state (Spencer et€al., 2009), but the actively ing through conduits formed by the intrusion
erupting sulci show the presence of organic of dikes. Although this scenario has been used
compounds and carbon dioxide (Brown et€ al., as the basis for interpretations on Titan in par-
2006), presumably derived from the inter- ticular (e.g., Lopes et€al., 2007), there are physical
ior. Larger grain sizes (up to 100 μm) of crys- difficulties with this mechanism because the
talline water ice are located near the sulci density of all proposed unvesiculated cryomag-
(Brown et€al., 2006; Porco et€al., 2006), possibly mas is greater than that of water-ice (specific-
due to preferential fallout of larger, slower ally Ice-I, the inferred bulk composition of icy
plume particles near their sources (Jaumann satellite crusts), and so watery cryomagmas are
et€ al., 2008, Hedman et€ al., 2009), or to sinter- negatively buoyant.
ing and growth of grains resulting from depos- There are a number of mechanisms that have
ition of plume gases or from the enhanced been proposed to overcome ascent problems
z depth (m) Carr, M.╯H. (2006). The Surface of Mars. New York, NY:
ρa atmospheric density (kg m−3) Cambridge University Press.
ρg gas density (kg m−3) Christensen, P.╯R., McSween, H.╯Y., Bandfield, J.╯L.
ρm magma density (kg m−3) et€al. (2005). Evidence for magmatic evolution and
ρ r country rock density (kg m−3) diversity on Mars from infrared observations.
Nature, 436, 504–882.
Choukroun, M., Grasset, O., Tobie, G. and Sotin, C.
References (2010). Stability of methane clathrate hydrates
under pressure: Influence on outgassing processes
Arvidson, R.╯E., Bonitz, R.╯G., Robinson, M.╯L. et€al. of methane on Titan. Icarus, 205, 581–593
(2009). Results from the Mars Phoenix lander Collins, G.╯C. and Goodman, J.╯C. (2007). Enceladus’
robotic arm experiment. Journal of Geophysical south polar sea. Icarus, 189, 72–82.
Research, 114, E00E02, doi:10.1029/2009JE003408. Crawford, G.╯D. and Stevenson, D.╯J. (1988).
Bandfield, J. (2002). Global mineral distribution of Gas-driven water volcanism and the resurfacing of
Mars. Journal of Geophysical Research 107(E6), 5042, Europa. Icarus, 73, 66–79.
doi:10.1029/2001JE001510. Croft, S.╯K., Lunine, J.╯I. and Kargel, J.╯S. (1988).
Bandfield, J.╯L., Hamilton, V.╯E. and Christensen, P.╯R. Equations of state of ammonia–water liquid:
(2000). A global view of martian surface composition Derivation and planetological applications. Icarus,
from MGS-TES. Science, 287, 1626–1630. 73, 279–293.
Barnes, J.╯W., Brown, R.╯H., Radebaugh, J. et€al. Croft, S.╯K., Kargel, J.╯S., Kirk, R.╯L. et€al. (1990). The
(2006). Cassini observations of flow-like features in geology of Triton. In Neptune and Triton, ed. D.╯P.
western Tui Regio, Titan. Geophysical Research Letters, Cruikshank. Tucson: University of Arizona Press,
33, L16204, doi:10.1029/2006GL026843. pp. 879–947.
Bleacher, J.╯E., Greeley, R., Williams, D.╯A., Cave, S.╯R. Crown, D.╯A. and Greeley, R. (1993). Volcanic
and Neukum, G. (2007). Trends in effusive style geology of Hadriaca Patera and the Eastern Hellas
at the Tharsis Montes, Mars, and implications Region of Mars. Journal of Geophysical Research, 98,
for the development of the Tharsis province. 3431–3451.
Journal of Geophysical Research, 112, E09005, Crumpler, L.╯S. and Aubele, J.╯C. (1978). Structural
doi:10.1029/2006JE002873. evolution of Arsia Mons, Pavonis Mons, and
Boynton, W.╯V., Feldman, W.╯C., Mitrofanov, I.╯G. et€al. Ascreus Mons: Tharsis Region of Mars. Icarus, 34,
(2004). The Mars Odyssey gamma-ray spectrometer 496–511.
instrument suite. Space Science Reviews, 110, 37–83. Davies, A., Keszthelyi, L., Williams, D. et€al. (2001).
Boynton, W.╯V., Taylor, G.╯J., Evans, L.╯G. et€al. Thermal signature, eruption style and eruption
(2007). Concentration of H, Si, Cl, K, Fe and Th evolution at Pele and Pillan on Io. Journal of
in the low- and mid-latitude regions of Mars. Geophysical Research, 106, 33,079–33,104.
Journal of Geophysical Research, 112, E12S99, Duxbury, N.╯S. and Brown, R.╯H. (1997). The role of
doi:10.1029/2007JE002887. an internal heat source for the eruptive plumes on
Brown, R. and Kirk, R. (1994). Coupling of volatile Triton. Icarus, 125, 83–93.
transport and internal heat flow on Triton. Journal Fagents, S.╯A. (2003). Considerations for effusive
of Geophysical Research, 99, 1965–1981. cryovolcanism on Europa: The post-Galileo
Brown, R.╯H., Clark, R.╯N., Buratti, B.╯J. et€al. (2006). perspective. Journal of Geophysical Research, 108,
Composition and physical properties of Enceladus’ E125139, doi:10.1029/2003JE002128.
surface. Science, 311, 1425–1428. Fagents, S.╯A., Greeley, R., Sullivan, R.╯J. et€al. (2000).
Carlson, R., Smythe, W., Lopes- Gautier, R. et€al. Cryomagmatic mechanisms for the formation of
(1997). The distribution of sulfur dioxide and other Rhadamanthys Linea, triple band margins, and other
infrared absorbers on the surface of Io in 1997. low albedo features on Europa. Icarus, 144, 54–88.
Geophysical Research Letters, 24, 2474–2482. Fagents, S.╯A., Lanagan, P.╯D. and Greeley, R. (2002).
Carlson, R.╯W., Kargel, J.╯S., Doute, Soderblom, L.╯A. Rootless cones on Mars: A consequence of lava–
and Dalton, J.╯B. (2007). Io’s surface composition. In ground ice interaction. In Volcano–Ice Interaction on
Io After Galileo: A New View of Jupiter’s Volcanic Moon, Earth and Mars, ed. J.╯L. Smellie and M.╯G. Chapman.
ed. R.╯M. C. Lopes and J.╯R. Spencer. Chichester, UK: Geological Society of London Special Publication,
Praxis-Springer, pp. 193–229. 202, pp. 295–317.
Fassett, C.╯I. and Head, J.╯W. (2006). Valleys on Hamilton, V.╯E., Christensen, P.╯R., McSween, H.╯Y.
Hecates Tholus, Mars: origin by basal melting of and Bandfield, J.╯L. (2003). Searching for the
summit snowpack. Planetary and Space Science, 54, source regions of martian meteorites using MGS
370–378. TES: Integrating martian meteorites into the
Fink, J.╯H. (1980). Surface folding and viscosity of global distribution of igneous materials on Mars.
rhyolite flows. Geology, 8, 250–254. Meteoritics and Planetary Science, 38(6), 871–885.
Fink, J.╯H. and Griffiths, R.╯W. (1990). Radial Hansen, C.╯J., Esposito, L., Stewart, A.╯I. F. et€al.
spreading of viscous gravity currents with (2006). Enceladus’ water vapor plume. Science, 311,
solidifying crust. Journal of Fluid Mechanics, 221, 1422–1425.
485–509. Hansen, C.╯J., Shemansky, D.╯E., Esposito, L.╯W. et€al.
Fink, J.╯H., Bridges, N.╯T. and Grimm, R.╯E. (1993). (2011). The composition and structure of the
Shapes of Venusian “pancake” domes imply Enceladus plume. Geophysical Research Letters, 38,
episodic emplacement and silicic composition. L11202, doi:10.1029/2011GL047415.
Geophysical Research Letters, 20, 261–264. Hartmann, W.╯K. and Berman, D.╯C. (2000). Elysium
Geissler, P.╯E. and Goldstein, D.B. (2007). Plumes Planitia lava flows: crater count chronology and
and their deposits. In Io After Galileo: A New View geological implications. Journal of Geophysical
of Jupiter’s Volcanic Moon, ed. R.╯M. C. Lopes and Research, 105, 15,011–015,025.
J.╯R. Spencer. Chichester, UK: Praxis-Springer, pp. Hauber, E., Bleacher, J. Gwinner, K., Williams,
163–192. D. and Greeley, R. (2009). The topography and
Gioia, G., Pinaki, C., Marshak, S. and Kieffer, S.╯W. morphology of low shields and associated
(2007). Unified model of tectonics and heat landforms of plains volcanism in the Tharsis
transport in a frigid Enceladus. Proceedings of the region on Mars. Journal of Volcanology and Geothermal
National Academy of Sciences, 103(34), 13 578–13 581. Research, 185, 69–95.
Glaze, L.╯S. and Baloga, S.╯M. (2002). Volcanic plume Head, J.╯W. and Pappalardo, R.╯T. (1999). Brine
heights on Mars: Limits of validity for convective mobilization during lithospheric heating on
models. Journal of Geophysical Research, 107, Europa: Implications for formation of chaos
doi:10.1029/2001JE001830. terrain, lenticula texture and color variations.
Greeley, R. and Crown, D.╯A. (1990). Volcanic geology Journal of Geophysical Research, 104, 27 143–27 155.
of Tyrrhena Patera, Mars. Journal of Geophysical Head, J.╯W. and Wilson, L. (1986). Volcanic processes
Research, 95, 7133–7149. and landforms on Venus: Theory, prediction and
Greeley, R. and Spudis, P.╯D. (1981). Volcanism on observations. Journal of Geophysical Research, 91,
Mars. Reviews of Geophysics and Space Physics, 19, 9407–9466.
13–41. Head, J.╯W., Chapman, C.╯R., Strom, R.╯G. et€al. (2011).
Greeley, R., Theilig, E. and Christensen, P. (1984). The Flood volcanism in the Northern High Latitudes
Mauna Loa sulfur flow as an analog to secondary of Mercury Revealed by MESSENGER. Science, 333,
(?) sulfur flows on Io. Icarus, 60, 189–199. 1853–1856.
Greeley, R., Bridges, N.╯T., Crown, D.╯A. et€al. (2000). Hedman, M.╯M., Nicholson, P.╯D., Showalter, M.╯R.
Volcanism on the red planet: Mars. In Environmental et€al. (2009). Spectral observations of the Enceladus
Effects on Volcanic Eruptions: From Deep Oceans to Deep plume with Cassini-VIMS. Astronomical Journal, 693,
Space, ed. J.╯R. Zimbelman and T.╯K. P. Gregg. New 1749–1762.
York: Kluwer/Plenum, 75–112. Hon, K., Kauahikaua, J., Denlinger, R. and
Greenberg, R., Hoppa, G., Tufts, B.╯R. et€al. (1999). Mackay, K. (1994). Emplacement and inflation
Chaos on Europa. Icarus, 141, 263–286. of pahoehoe sheet flows: Observations and
Gregg, T.╯K. P. and Lopes R.╯M. (2008). Lava lakes on measurements of active lava flows on Kilauea
Io: New perspectives from modeling. Icarus, 194, Volcano, Hawaii. Geological Society of America
166–172, doi:10.1016/j.icarus.2007.08.042. Bulletin, 106, 351–370.
Grosfils, E.╯B., Aubele, J., Crumpler, L., Gregg, T.╯K. Howell, R.╯R. and R.╯M. C. Lopes (2011). Morphology,
P. and Sakimoto, S. (2000). Volcanism n Earth’s temperature, and eruption dynamics at Pele. Icarus,
seafloor and Venus. In Environmental Effects of 213. 593–607, doi:10.1016/j.icarus.2011.03.008.
Volcanic Eruptions: From Deep Oceans to Deep Space, Hulme, G. (1974). The interpretation of lava flow
ed. J.╯R. Zimbelman and T.╯K. P. Gregg. New York: morphology. Geophysical Journal of the Royal
Kluwer/Plenum, pp. 113–142. Astronomical Society, 39(2), 361–383.
Hulme, G. (1976). The determination of the Kirk, R.╯L., Howington-Kraus, E., Barnes, J.╯W. et€al.
rheological properties and effusion rate of an (2010). La Sotra y las otras: Topographic evidence
Olympus Mons lava. Icarus, 27(2), 207–213. for (and against) cryovolcanism on Titan. Presented
Jaeger, W.╯L., Turtle, E.╯P., Keszthelyi, L.╯P. et€al. (2003). at the American Geophysical Union Fall Meeting,
Orogenic tectonism on Io. Journal of Geophysical 2010, San Francisco.
Research, 108, doi:10.1029/2002JE001946. Kivelson, M.╯G., Khurana, K.╯K., Russell, C.╯T. et€al.
Jaeger, W.╯L., Keszthelyi, L.╯P., McEwen, A.╯S., Dundas, (2000). Galileo magnetometer measurements: A
C.╯M. and Russell, P.╯S. (2007). Athabasca Valles, stronger case for a subsurface ocean at Europa.
Mars: A lava-draped channel system. Science, 317, Science, 289, 1340–1343.
1709–1711. Leone, G. and Wilson, L. (2001). The density structure
Jaumann, R., Stephan, K., Hansen, G.╯B. et€al. of Io and the migration of magma through its
(2008). Distribution of icy particles across lithosphere. Journal of Geophysical Research, 106, 32
Enceladus’ surface as derived from Cassini-VIMS 983–32 995.
measurements. Icarus, 193, 407–419. Lopes, R. and Williams, D. (2005). Io after Galileo.
Jessup, K.╯L., Spencer, J.╯R., Ballester, G.╯E. et€al. (2004). Reports on Progress in Physics, Institute of Physics
The atmospheric signature of Io’s Prometheus Publishing, 68, 303–340.
plume and anti-jovian hemisphere: Evidence for a Lopes, R., Kamp, L.╯W., Douté, S. et€al. (2001). Io in the
sublimation atmosphere. Icarus, 169, 197–215. near-infrared: NIMS results from the Galileo fly-bys
Jewitt, D.╯C. and Luu, J. (2004). Crystalline water ice in 1999 and 2000. Journal of Geophysical Research,
in Kuiper Belt Object (50000) Quaoar. Nature, 432, 106, 33 053–33 078.
731–733. Lopes, R., Kamp, L.╯W., Smythe, W. et€al. (2004). Lava
Jolliff, B.╯K., Wiseman, S.╯A., Lawrence, S.╯J. et€al. lakes on Io. Observations of Io’s volcanic activity
(2011). Non-mare silicic volcanism on the lunar from Galileo during the 2001 fly-bys. Icarus, 169,
farside at Compton–Belkovich. Nature Geosciences, 4, 140–174.
566–571. Lopes, R.╯M. C., Mitchell, K.╯L., Stofan, E.╯R. et€al.
Kargel, J.╯S., Croft, S.╯K., Lunine, J.╯I. and Lewis, J.╯S. (2007). Cryovolcanic features on Titan’s surface as
(1991). Rheological properties of ammonia–water revealed by the Cassini Titan Radar Mapper. Icarus,
liquids and crystal–liquid slurries: Planetological 186, 395–412.
applications. Icarus, 89, 93–112. Lopes, R.╯M. C., Stofan, E.╯R., Peckyno, R. et€al.
Keszthelyi, L., McEwen, A.╯S., Phillips, C.╯B. et€al. (2010a). Distribution and interplay of geologic
(2001). Imaging of volcanic activity on Jupiter’s processes on Titan from Cassini RADAR data.
moon Io by Galileo during GEM and GMM. Journal Icarus, 205, 540–588.
of Geophysical Research, 106, 33,025–33,052. Lopes, R.╯M.C., Mitchell, K.L., Williams, D.A.,
Keszthelyi, L.╯P., Jaeger, W. and Milazzo, M. et€al. and Mitri, G. (2010b): Beyond Earth: How
(2007). New estimates for Io eruption temperatures: extra-terrestrial volcanism has changed our
implications for the interior. Icarus, 192, 491–502. definition of a volcano. In: “What’s a volcano? New
Khurana, K.K. ; X. Jia ; M.G. Kivelson; F. Nimmo ; answers to an old question”. Geological Society of
G. Schubert ; C.T. Russell (2011). Evidence of a America Special Paper 470 (Eds. E. Canon and A.
global magma ocean in Io’s interior. Science, 332, Szakacs). Pp. 11–30.
1186–1189. Lorenz, R.╯D. (1996). Pillow lava on Titan:
Kieffer, S.╯W. (1982). Dynamics and thermodynamics expectations and constraints on cryovolcanic
of volcanic eruptions: Implications for the processes. Planetary and Space Science, 44,
plumes on Io. In Satellites of Jupiter, ed. D. 1021–1028.
Morrison. Tuscon, AZ: University of Arizona Press, Macdonald, G.╯A. (1972). Volcanoes. New Jersey:
pp.€647–723. Prentice Hall.
Kieffer, S.╯W., Lopes-Gautier, R., McEwen, S. et€al. Manga, M. and Wang, C.╯Y. (2007). Pressurized oceans
(2000). Prometheus: Io’s wandering plume. Science, and the eruption of liquid water on Europa and
288, 1204–1208. Enceladus. Geophysical Research Letters, 34, L07202,
Kieffer, S.╯W., Lu, X., Bethke, C.╯M. et€al. (2006). A doi:10.1029/2007GL029297.
clathrate reservoir hypothesis for Enceladus’ south Marchis, F., Prangé, R. and Christou, J. (2000).
polar plume. Science, 314, 1764–1766, doi:10.1126/ Adaptive optics mapping of Io’s volcanism in the
science.1133519. IR (3.8 μm). Icarus, 148, 384–396.
Mastin, L.╯G. (1995). Thermodynamics of gas and Mouginis-Mark, P.╯J., Wilson, L. and Zimbelman, J.╯R.
steam-blast eruptions. Bulletin of Volcanology, 57, (1988). Polygenic eruptions on Alba Patera, Mars.
85–98. Bulletin of Volcanology, 50, 361–379.
Mastin, L.╯G. (2005). The controlling effect of viscous Nelson, R.╯M., Kamp, L.W., Matson, D.L. et€al. (2009a).
dissipation on magma flow in silicic conduits. Saturn’s Titan: Surface change, ammonia, and
Journal of Volcanology and Geothermal Research, 143, implications for atmospheric and tectonic activity,
17–28. Icarus, 199, 429–441.
McEwen, A.╯S. and Soderblom, L. (1983). Two classes Nelson, R.╯M., Kamp, L.╯W., Lopes, R.╯M. C. et€al.
of volcanic plumes on Io. Icarus, 58, 197–226. (2009b). Photometric changes on Saturn’s
McEwen, A.╯S., Keszthelyi, L.╯P., Spencer, J. et€al. moon Titan: Evidence for cryovolcanism.
(1998). High-temperature silicate volcanism on Geophysical Research Letters, 36, L04202,
Jupiter’s moon Io. Science, 281, 87–90. doi:10.1029/2008GL036206.
McKinnon, W.╯B. (1984). On the origin of Triton and Neukum, G., Jaumann, R., Hoffmann, H. et€al. (2004).
Pluto. Nature, 311, 355–358. Recent and episodic volcanic and glacial activity
McSween, H.╯Y. (1994). What we have learned about on Mars revealed by the High Resolution Stereo
Mars from SNC Meteorites. Meteoritics, 29, 757–779. Camera. Nature, 432, 971–979.
McSween, H.╯Y., Murchie, S.╯L., Crisp, J.╯A. et€al. (1999). Newman, S.╯F., Buratti, B.╯J., Brown, R.╯H. et€al.
Chemical, multispectral, and textural constraints (2008). Photometric and spectral analysis
on the composition and origin of rocks at the of the distribution of crystalline and amorphous
Mars Pathfinder landing site. Journal of Geophysical ices on Enceladus as seen by Cassini. Icarus 193,
Research, 104, 8679–8715. 397–406.
McSween, H.╯Y., Taylor, G.╯J. and Wyatt, M.╯B. (2009). Niemann, H.╯B., Atreya, S.╯K., Bauer, S.╯J. et€al.
Elemental composition of the martian crust. (2005). The abundances of constituents of Titan’s
Science, 324, 736–739. atmosphere from the GCMS instrument on the
Milazzo, M., Keszthelyi, L. and McEwen, A. (2001). Huygens probe. Nature, 438, 778–794.
Observations and initial modeling of lava–SO2 Nimmo, F. and Pappalardo, R.╯T. (2006).
interactions at Prometheus, Io. Journal of Geophysical Diapir-induced reorientation of Enceladus. Nature,
Research, 106, 33 121–33 127. 441, 614–616, 2006.
Mitchell, K.╯L. (2005). Coupled conduit flow and Nyquist, L.╯E., Bogard, D.╯D., Shih, C.-Y. et€al.
shape in explosive volcanic eruptions. Journal of (2001). Ages and geologic histories of martian
Volcanology and Geothermal Research, 143, 187–203. meteorites. In Chronology and Evolution of Mars, ed. R.
Mitri, G. and Showman, A.╯P. (2008a). A model for Kallenbach, J. Geiss and W.╯K. Hartmann. Boston,
the temperature-dependence of tidal dissipation in MA: Kluwer Academic Publishers, pp. 105–164.
convective plumes on icy satellites: implications Pappalardo, R.╯T., Belton, M.╯J. S., Breneman, H.╯H.
for Europa and Enceladus. Icarus, 195, 758–764. et€al. (1999). Does Europa have a subsurface ocean?
Mitri, G. and Showman A.╯P. (2008b). Thermal Evaluation of the geological evidence. Journal of
convection in ice-I shells of Titan and Enceladus. Geophysical Research, 104, 24 015–024 055.
Icarus, 193, 387–396. Pearl, J., Hanel, R., Kunde, V. et€al. (1979).
Mitri, G., Showman, A.╯P., Lunine, J.╯I. and Identification of gaseous SO2 and new upper limits
Lopes, R.╯M. C. (2008). Resurfacing of Titan by for other gases on Io. Nature, 280, 755–758.
ammonia-water cryomagma. Icarus, 196, 216–224. Porco, C.╯C., Helfenstein, P., Thomas, P.╯C. et€al.
Moore, W.╯B. (2001). The thermal state of Io. Icarus, (2006). Cassini observes the active south pole of
154, 548–550. Enceladus. Science, 311, 1393–1401.
Moore, J.╯M. and Pappalardo, R.╯T. (2011). Titan: An Postberg, F., Kempf, S., Schmidt, J. et€al. (2009).
exogenic world? Icarus, 212, 790–806. Sodium salts in E-ring ice grains from an ocean
Morabito, L.╯A., Synnott, S.╯P., Kupferman, P.╯N. and below the surface of Enceladus. Nature, 459,
Collins, S.╯A. (1979). Discovery of currently active 1098–1101, doi: 10.1038/nature08046.
extraterrestrial volcanism. Science, 204, 972. Postberg, F., Schmidt, J. Hillier, J., Kempf, S. and
Mouginis-Mark, P.╯J., Wilson, L. and Head, J.╯W. Srama, R. (2011). A salt-water reservoir as the
(1982). Explosive volcanism on Hecates Tholus, source of a compositionally stratified plume on
Mars: Investigation of eruption conditions. Journal Enceladus. Nature, 474, 620–622, doi:10.1038/
of Geophysical Research, 87, 9890–9904. nature10175.
Poulet, F., Gomez, C., Bibring, J.-P. et€al. (2007). Spencer, J.╯R., Pearl, J.╯C., Segura, M. et€al. (2006).
Martian surface mineralogy from OMEGA/MEx: Cassini encounters Enceladus: Background and the
Global mineral maps. Journal of Geophysical Research, discovery of a south polar hot spot. Science, 311,
112, E08S02, doi:10.1029/2006JE002840. 1401–1405.
Radebaugh, J., Keszthelyi, L.╯P., McEwen, A.╯S. et€al. Spencer, J.╯R., Stern, S.╯A., Cheng, A.╯F. et€al. (2007).
(2001). Paterae on Io: A new type of volcanic Io volcanism seen by New Horizons: A major
caldera? Journal of Geophysical Research, 106, 33 eruption of the Tvashtar volcano. Science, 318,
005–33020. 240–243.
Radebaugh, J., McEwen, A.╯S., Milazzo, M.╯P. et€al. Spencer, J.╯R., Barr, A.╯C., Esposito, L.╯W. et€al. (2009).
(2004). Observations and temperatures of Io’s Pele Enceladus: An active cryovolcanic satellite. In
Patera from Cassini and Galileo spacecraft images. Saturn From Cassini-Huygens, ed. M.╯K. Dougherty,
Icarus, 169, 65–79. L.╯W. Esposito and S.╯M. Krimigis. Dordrecht,
Rathbun, J.╯A., Spencer, J.╯R., Davies, A.╯G., Howell, Netherlands: Springer, pp. 683–724.
R.╯R. and Wilson, L. (2002). Loki, Io: A periodic Spitale, J.╯N. and Porco, C.╯C. (2007). Association of
volcano. Geophysical Research Letters, 29, 1443, the jets of Enceladus with the warmest regions on
doi:10.1029/2002GL014747. its south-polar fractures. Nature, 449, 695–697.
Ruff, S.╯W., Christensen, P.╯R., Blaney, D.╯L. et€al. Squyres, S.╯W., Arvidson, R.╯E., Bollen, D. et€al. (2006).
(2006). The rocks of Gusev crater as viewed by the Overview of the Opportunity Mars Exploration
Mini-TES instrument. Journal of Geophysical Research, Rover Mission to Meridiani Planum: Eagle crater to
111, E12S18, doi:10.1029/2006JE002747. Purgatory ripple. Journal of Geophysical Research, 111,
Schaber, G.╯G. (1980). The surface of Io: Geologic E12S12, doi:10.1029/2006JE002771.
units, morphology, and tectonics. Icarus, 43, Stofan, E.╯R. and Smrekar, S.╯E. (2005). Large
302–333. topographic rises, coronae, large flow fields and
Showman, A. P, Mosqueira, I. and Head, J.╯W. (2004). large volcanoes on Venus: Evidence for mantle
On the resurfacing of Ganymede by liquid-water plumes? In Plates, Plumes, and Paradigms, ed. G.╯R.
volcanism. Icarus, 172, 625–640. Foulger, J.╯H. Natland, D.╯C. Presnall and D.╯L.
Skok, J.╯R., Mustard, J.╯F., Ehlmann, B.╯L., Milliken, Anderson. Geological Society of America Special
R.╯E. and Murchie, S.╯L. (2010). Silica deposits in Paper, 388, pp. 841–861.
the Nili Patera caldera on the Syrtis Major volcanic Tobie G., Ä„adek O. and Sotin, C. (2008). Solid tidal
complex on Mars. Nature Geoscience, doi:10.1038/ friction above a liquid water reservoir as the origin
ngeo990. of the south pole hotspot on Enceladus. Icarus,
Smith, B.╯A., Soderblom, L.╯A., Banfield, D. et€al. 196, 642–652
(1989). Voyager 2 at Neptune: Imaging science Treiman, A.╯H. (2007). Geochemistry of Venus’s
results. Science, 246, 1422–1449. surface: Current limitations as future
Soderblom, L., Kieffer, S.╯W., Becker, T.╯L. et€al. (1990). opportunities. In Exploring Venus as a Terrestrial
Triton’s geyser-like plumes: Discovery and basic Planet. Washington, DC: American Geophysical
characterization. Science, 250, 410–415. Union Monograph 176, pp. 7–22.
Soderblom, L.╯A., Brown, R.╯H., Soderblom, J.╯M. Treiman, A.╯H., Gleason, J.╯D. and Bogard, D.╯D.
et€al. (2009). The geology of Hotei Regio, Titan: (2000). The SNC meteorites are from Mars.
Correlation of Cassini VIMS and RADAR. Icarus, Planetary and Space Science, 48, 1213–1230.
204, 610–618. Veeder, G.╯J., Matson, D.╯L., Johnson, T.╯V., Blaney,
Sotin, C., Jaunmann, R., Buratti, B.╯J. et€al. (2005). D.╯L. and Goguen, J.╯D. (1994). Io’s heat flow
Release of volatiles from a possible cryovolcano from infrared radiometry: 1983–1993. Journal of
from near-infrared imaging of Titan. Nature, 435, Geophysical Research, 99, 17 095–17 162.
786–789. Wadge, G. and Lopes, R.╯M. C. (1991). The lobes of
Spahn, F., Schmidt, J., Albers, N. et€al. (2006). lava flows on Earth and Olympus Mons, Mars.
Cassini dust measurements at Enceladus and Bulletin of Volcanology, 54, 10–24.
implications for the origin of the E ring. Science, Wall, S.╯D., Lopes, R.╯M., Stofan, E.╯R. et€al. (2009).
311, 1416–1418. Cassini RADAR images at Hotei Arcus and Western
Spencer, J.╯R., Jessup, K.╯L., McGrath, M.╯A., Ballester, Xanadu, Titan: Evidence for recent cryovolcanic
G.╯E. and Yelle, R. (2000). Discovery of gaseous S2 activity. Geophysical Research Letters, 36, L04203,
in Io’s Pele plume. Science, 288, 1208–1210. doi:1029/2008GL036415.
Waite, J.╯H., Combi, M.╯R., Ip, W.-H. et€al. (2006). Wilson, L. and Head, J.╯W. (1994). Mars: Review
Cassini Ion and Neutral Mass Spectrometer: and analysis of volcanic eruption theory and
Enceladus plume composition and structure. relationships to observed landforms. Reviews of
Science, 311, 1419–1422. Geophysics, 32, 221–264.
Waite, J.╯H., Lewis, W.╯S., Magee, B.╯A. et€al. (2009). Wilson, L. and Head, J.╯W. (2001). Lava fountains
Liquid water on Enceladus from observations of from the 1999 Tvashtar Catena fissure eruption
ammonia and 40Ar in the plume. Nature, 460, on Io: Implications for dike emplacement
487–490. mechanisms, eruption rates and crustal structure.
Warner, N. and Gregg, T.╯K. P. (2003). Evolved lavas Journal of Geophysical Research, 106, 32 997–33 004.
on Mars? Observations from southwest Arsia Mons Wilson, L. and Head, J.╯W. (2004). Evidence for a
and Sabancaya volcano, Peru. Journal of Geophysical massive phreatomagmatic eruption in the initial
Research, 108, doi:10.1029/2002JE001969. stages of formation of the Mangala Valles outflow
Weitz, C.╯M. and Head, J.╯W. (1999). Spectral channel, Mars. Geophysical Research Letters, 31,
properties of the Marius Hills volcanic complex and L15701, doi:10.1029/2004GL020322.
implications for the formation of lunar domes and Wilson, L. and Head, J.╯W. (2007). Explosive volcanic
cones. Journal of Geophysical Research, 104, 18 933–18 eruptions on Mars: Tephra and accretionary
956. lapilli formation, dispersal and recognition in the
Wichura, H., Bousquet, R. and Oberhansli, R. (2010). geologic record. Journal of Volcanology and Geothermal
Emplacement of the mid-Miocene Yatta lava flow, Research, 163, 83–97.
Kenya: Implications for modeling long channeled Wilson, L. and Mouginis-Mark, P.╯J. (2003a).
lava flows. Journal of Volcanology and Geothermal Phreatomagmatic explosive origin of Hrad
Research, 198, 325–338. Vallis, Mars. Journal of Geophysical Research, 108,
Williams, D.╯A. and Greeley, R. (1994). Assessment doi:10.10292002JE001927.
of antipodal-impact terrains on Mars. Icarus, 110, Wilson, L. and Mouginis-Mark, P.╯J. (2003b).
196–202. Phreatomagmatic dike–cryosphere interactions as
Williams, D.╯A. and Howell, R.╯R. (2007). Active the origin of small ridges north of Olympus Mons,
volcanism: Effusive eruptions. In Io after Galileo, ed. Mars. Icarus, 165, 242–252.
R.╯M. C. Lopes and J.╯R. Spencer. Chichester, UK: Zhang, J., Goldstein, D.╯B., Varghese, P.╯L. et€al. (2003).
Praxis, pp. 133–161. Simulation of gas dynamics and radiation in
Williams, D.╯A., Greeley, R. and Davies, A.╯G. (2001). volcanic plumes on Io. Icarus, 163, 182–197.
Evaluation of sulfur flow emplacement on Io from Zimbelman, J.╯R. (1985). Estimates of rheologic
Galileo data and numerical modeling. Journal of properties for flows on the Martian volcano
Geophysical Research, 106, 33 161–133 174. Ascraeus Mons. Proceedings of the 16th Lunar and
Williams, D.╯A., Turtle, E.╯P., Keszthelyi, L.╯P. et€al. Planetary Science Conference, Journal of Geophysical
(2004). Geologic mapping of the Culann-Tohil region Research, 90 (supplement), 157–162.
of Io from Galileo imaging data. Icarus, 169, 80–97. Zimbelman, J.╯R. and Gregg, T.╯K. P. (2000). Volcanic
Williams, D.╯A., Greeley, R. Fergason, R.╯L. et€al. diversity throughout the Solar System. In
(2009). The Circum-Hellas Volcanic Province: Environmental Effects on Volcanic Eruptions: From Deep
Overview. Planetary and Space Science, 57, 895–916. Oceans to Deep Space, ed. J.╯R. Zimbelman and T.╯K. P.
Williams, D.╯A., Keszthelyi, L.╯P., Crown, D.╯A. et€al. Gregg. New York: Kluwer/Plenum, pp. 1–8.
(2011). Volcanism on Io: New insights from global
geologic mapping. Icarus, 214, 91–112. Online resources available at www.cambridge.
Wilson, L. and Head, J.╯W. (1981). Ascent and org/fagents
eruption of basaltic magma on the Earth and
Moon. Journal of Geophysical Research, 86, 2971–3001. • Links to websites of interest
Wilson, L. and Head, J.╯W. (1983). A comparison of • Exercises
volcanic eruption processes on Earth, Moon, Mars, • Answers to exercises
Io and Venus. Nature, 302, 663–669.
‘a’a, 85, 92 boiling liquid expanding vapor cohesion, 217, 303, 311
accretion, deposit, 302, 303, 307 explosion (BLEVE), 369 cohesive flow. See€lahar, cohesive
accretionary lapilli, 184, 234 bomb, 131, 145, 168, 233, 268 coignimbrite plume, 154, 166, 167,
acoustic Bond number, 242 168, 184
data, 118, 122, 304, 375 boundary integral method, 351 column, 130, 153, 156, 174, 233
pressure, 360, 361, 362, 372, 378 Boussinesq assumption, 18, 142, 316 collapse, 156, 161, 163, 165, 203
adiabatic expansion, 116, 120, 138, box model, 208, 209, 210, 211 height, 153, 154, 156, 158, 162,
248, 249 breadcrust texture, 131 176, 178, 181, 183, 186, 190,
advection, 7, 12, 62, 333, 339, 341, brittle deformation, 64, 75, 134 192
See€lava flow, heat sources, Brunt-Väisälä frequency, 142, 158 Mars, 396
advection bubbles Mars, 387
wind, 175, 180, 181, 182 breakup, 62 vulcanian, 130
advection-diffusion relation, 317 coalescence, 61–62, 74, 110, 111, columnar jointing, 288
Advection-Diffusion-Sedimentation 112, 121, 123 compressibility, 42, 56, 70, 72,
(ADS) models, 179–83, 188–91 collapse, 132 140–41, 361
limitations, 183–86 expansion, 60, 110, 112, 134, 135, compressible flow, 220, 360
one-dimensional, 180–81 396 computational fluid dynamics, 117
three-dimensional, 181–83 growth, 60, 68, 69, 71, 110 concentration
two-dimensional, 181 diffusion-limited, 61 of particles, 179, 204, 205, 209,
aerosols, 153, 169 solubility-limited, 61 211, 212, 312, See€sediment
Afar, Ethiopia, 35 viscosity-limited, 60–61, 73, 132 concentration
aggradation, 205, 219 nucleation, 59–60, 110, 396 conduction, 9, 13, 42, 163, 291,
aggregation, of particles, 174, 181, heterogeneous, 59 See€lava flow, heat loss,
183, 184 homogeneous, 59 conduction
andesite, 179, 287, 292, 388, 391, number density, 60, 145, 146 conduit
392 rise speed, 62–63, 110, 113 erosion of, 56
annular flow, 63 volume fraction, 64, 68, 69, 113, geometry, 56, 112, 113, 121, 122,
apophyses, 288 135, See€vesicularity 345, 351
Arrhenian relationship, 90 bubbly flow, 68, 110, 111, inclined, 113, 115
ascent velocity See€dispersed flow conservation equations
bubble. See€bubbles, rise speed bulk modulus, 15, 237, 251, 348, 349 buoyancy, 142
gas slug, 114–15, See€gas slug, rise bulking, 306, 318 energy, 9, 19, 70, 137, 160
speed buoyancy mass, 19, 60, 69, 71, 137, 142, 159,
magma, 111, 134, See€rise speed, flux, 142, 157, 158, 162 179, 212, 220, 313, 315, 317
magma force, 46, 141, 207, 395, 397 momentum, 19, 39, 69, 71, 137,
Mars, 385 buoyant plume theory (BPT), 179, 142, 160, 207–08, 220, 313,
atmospheric boundary layer (ABL), 191 317
181 burst convection, 13, 244, See€lava flow,
autobreccia, 290, 292 strombolian, 108, 373, See€gas heat loss, convection
avalanche slug, burst in magma chambers, 18–22
debris, 222, 302, 321 Rayleigh-Bénard, 18
dry, 316, 318 caldera, 6, 17, 203, 264, 389, 398 solid state, 406
rock, 204, 310 CALMET, 182, 184, 190 conversion ratio, 248, 249
Avrami equation, 65 Capelinhos, 231 cooling-limited flow, 95, 99–101
Capillary number (Ca), 62, 67 Coulomb friction, 213, 214, 216,
basaltic andesite, 107, 268, 287, 388 capillary pressure, 59, 62 311, 317, 318
bedding, 233, 281, 302 cellular automata model, 100–01, Coulomb mixture model, 316–18,
bedload, 302, 307 220 321
Bernoulli’s equation, 351, 361 Chapman-Jouguet (C-J) condition, crack
Bessel function, 374, 376 241, 242 stiffness, 349
Bingham fluid, 86, 87, 88, 98, 222, Chézy coefficient, 314, 320 wave, 349, 352, 375
311 choked flow, 70, 351 crater lake, 165, 302, 316, 319, 322
Biot number (Bi), 9, 24 CO2, 110, 121, 268, 278, 279, 283, critical point, 237, 238, 239, 246,
block, ballistic, 131, 145, 168, 233 See€volatiles, CO2 content 247
block-and-ash flow, 184, 205 coda, 347, 367 cryomagma, 402, 406
fall deposit, 107, 168, 233, 276, 389 stress criterion, 68, 135 inertial, 207, 208, 212, 219
hawaiian, 107, 109 passive, 244 gray-body, 92
plinian, 107, 154 phreatomagmatic, 134, 157 Green’s function, 333, 340, 341
strombolian, 107, 109 subglacial, 289 ground deformation, 7, 33
supraglacial, 276, 291, 292 threshold, 136 groundwater, 129, 133, 231, 233,
volume, 177 timescale, 146 398
vulcanian, 130 wave, 129, 130, 134, 136, 138,
fall speed, 164, See€settling velocity 146, 283 H2O, 110, 278, 283, See€volatiles,
fall time threshold (FTT), 190 velocity, 136 H2O content
FALL3D, 182–83, 190, 192 free surface hawaiian activity, 74, 109, 111,
fallout. See€sedimentation lava, 284 112, 153, 168–70, 268, 287,
fault slip, 334 magma, 108, 112, 135 See€Chapter 8
film drainage, 62, 339, 340 free-surface flow, 117 hazard
finger jets, 144 friction coefficient, 213, 222 assessment, 173, 174, 176, 177,
finite difference method, 117, 182, Froude number (Fr), 114, 115, 208, 184, 195, 206, 221, 223, 301,
220, 341, 351 210, 218 321
finite element method, 220 fuel–coolant interaction (FCI), 235, lava, 86, 101
finite volume method, 220, 322 286 real-time, 195, 222
firn, 278, 279 aviation, 191
fissure eruption, 56, 169, 266, 284, Gakkel Ridge, 262, 268 map, 308, 321
287, 366 Galeras, 351 Hazen’s law, 209
FLO-2D, 312, 315 Ganymede, 402 heat conduction equation, 250
flood lava, 169, 260, 394 gas piston, 343, 366 Heimaey, 107
flow resistance, 308, 312, 314, 321 gas slug, 74, 109, 110, 111, 339, 343, Hekla, 332
Flow-3D®, 117 347, 351 Herschel-Bulkley equation, 311
FLOWFRONT, 98, 100, 101 ascent, 113–18 homogeneous flow, 62, 69–70, 71,
FLOWGO, 96, 100 burst, 118–20, 123, 339, 366, 373 74, 111
fluid-filled crack, 332, 347, 349, 351, expansion, 115–17 homogeneous nucleation
365 formation, 110–13 temperature (HNT), 237, 247
fluidization, 216 foam collapse model, 110, 112 hot spot, 259
minimum velocity, 216 rise speed model, 110, 111–12 Hugoniot, 241
velocity, minimum, 217 oscillations, 118 hyaloclastite, 232, 233, 267, 281,
foam, magmatic, 20, 110, 268, 269, rise speed, 123 287, 288
278, 285, 370 gas–particle mixture, 130, 141, 155, hyalotuff, 232
collapse, 112, 278 159, 203, 213, 216, 370 hydraulic jump, 312, 316
Forcheimer’s equation, 64, 69 gas-thrust region, 155, 174, See€jet, hydraulic routing model, 312, 316,
fountaining activity, 109, 111, 112, momentum-driven 320
168–70, 231, 275, 290, 366 Gibbs function, 251, 252 fully dynamic, 314, 316
subglacial, 286 Gjálp, 276, 281 hydroclast, 233, 234
Fourier transform, 340, 362 glacier, 338, See€ice sheet hydrodynamic equivalence, 211
Fourier’s Law, 94 cold-based, 283, 292 hydrograph, 301, 303, 307, 312, 313,
fractionation warm-based, 283, 292 319, 321
vs. crustal melting, 12–14 glaciovolcanic sequence, 282 hydrologic model, 308, 313
fracture toughness, 35, 39, 41, 45, glass transition temperature, 239, hydromagmatism, 231,
46 266 See€phreatomagmatic
fragmentation, 67–68, 108, 111, 134, gliding, 367, 377 eruption
135–36, 168, 169, 231, 239, Gorda Ridge, 268 hydrothermal
252, 278, 285 grading, 221, 233, 281 alteration, 233, 234
brittle, 73, 134, 135, 237, 241, 244, normal, 211, 214 explosion, 232
251–52 reverse, 213, 214 fluid, 7, 332, 333
energy, 252 grainsize distribution, 175, 181, 183, seismic event, 350
hydrovolcanic. 233, 241, 194, 252, 302, 303, See€particle system, 248, 375
268, See€fragmentation, size distribution hydrovolcanic
phreatomagmatic granular flow, 206, 212–19, 222, deposits, 233
mechanisms, 176, 396 311, 316 interaction
bubble volume fraction dry, 213–16 non-explosive, 244, 249
criterion, 68, 135 gas–particle, 213 subsonic explosive, 244, 249
hydrodynamic, 68, 74, 236, 241 granular stress, 73 supersonic explosive, 245, 249,
inertial. See€hydrodynamic granulation, thermal, 233, 239, 244, 250
potential energy criterion, 68 247 products, 233–34
strain-rate criterion, 68, 134–35 gravity current, 174, 206–07 hydrovolcanism, 230, 231
hyperconcentrated flow, 288, 289, jökulhlaup, 232, 276, 285, 287, 288, height, 168
300, 302, 306, 307, 314, 319, 289 Io, 400
320 Juan de Fuca Ridge, 264, 266, 267, subglacial, 284
268 lake, 91, 118, 122, 366, 402
ice pillar, 263
cap. See€ice sheet Kasatochi, 368 tube, 85, 389, 394, 400
deformation, 285 Kelvin-Helmholtz instability, 210 types, 85
sheet Kilauea, 109, 168, 266, 340, 350, lava flow
thickness, 278, 286, 288 351, 362, 366 heat budget, 92–95
ideal gas, 15, 69, 136, 138, 140 kinematic heat loss
ignimbrite, 6, 184, 210, 217, 391, sieving, 213 conduction, 94, 266, 387, 388
See€pyroclastic density current wave, 118, 315 convection, 265, 387, 388, 397
(PDC) Kozeny-Carman equation, 64 forced convection, 93
incompressible flow, 19, 110, 313, Krakatau, 367 free convection, 93
316, 351, 360 radiation, 92, 387, 388, 397
inertial number (I), 213 La Soufrière, St. Vincent, 158 heat sources
inflation. See€lava flow, inflation lag deposit, 292 advection, 94, 266
infrasound, 109, 118, 123, Lagrangian model, 20, 23, 179 latent heat of crystallization, 94
See€Chapter 16 HYSPLIT, 182 inflated, 260, 400
definition, 359 NAME, 182 inflation, 263, 284
recording, 362–65 PUFF, 182 path prediction, 95–98
inundation area, 86, 308, 316, 318 VOL-CALPUFF, 182 rheomorphic, 168, 286
inversion methods, 180, 181, 183, lahar, 131, 222, 291, 292, subglacial, 277
189, 193, 194, 338 See€Chapter 14 submarine, 263–65
Io, 385 advance rate, 300 supraglacial, 290, 291
eruption cohesive, 303 Type I model, 86, 95
explosion-dominated, 400 deposition, 318 Type II model, 86, 89, 95, 101
flow-dominated, 400 deposits, 302, 303, 307 least principal stress, 15, 277
intra-patera, 402 discharge, 300, 308, 310 LEFM. See€linear elastic fracture
processes, 399–402 initiation, 301 mechanics
rate, 398 models, 307–20 Leidenfrost effect, 247
temperature, 399 non-cohesive, 303 levée, 95, 203, 214, 222, 303, 322
velocity, 400 primary, 302 limu o Pele, 268
exsolution, 400 rheology, 305, 307, 308, 310–13 linear elastic fracture mechanics
lava runout. See€runout, lahar (LEFM), 35, 37, 39
flow field, 398, 400 secondary, 302 linear elasticity equations, 16
lake, 402 stage, 308 lithics, 130, 174, 233
magma composition, 399 surge, 307, 312, 320 lithofacies
patera, 398 transformation, 303, 306, 319 dike collapse, 281, 282
plumes, 398, 400, 402 travel time, 310 lock-exchange experiments, 210
deposits, 398, 400 velocity, 317, 318 Loihi, 258, 268
sulfur flow, 399 volume, 300 long-period (LP) event, 331, 332, 347,
superheated magma, 399 LAHARZ, 308–10 365, 366, 368
ultramafic magma, 399 laminar flow, 64, 67, 116, 132, 317, lubrication theory, 38, 39
volcanic features, 398–99 395 lumped-parameter model, 351
isentropic expansion, 70, 248, 249 landslide, 222, 269, 307, 318, 338
isomass, 175, 189 lapilli, 130, 175, 203, 268 maar, 231, 234
isopach, 175, 176, 177 tuff, 288, 289 Mach number (M), 70, 139, 374, 376,
isopleth, 175, 177, 178 large eddy simulation (LES), 182, 378
isothermal expansion, 117, 136, 191, 193 MAGFLOW, 100
137 large igneous province (LIP), 260–61 magma
Izu peninsula, 37 latent heat, 9, 94, 162, 235 chamber
ice, 280 definition, 6, 56
Jeffreys equation, 87 lateral blast, 154, 203 Mars, 388, 395
jet, 109, 119, 120, 241 Lattice Boltzmann model, 117 pressure, 72, See€overpressure,
cypressoid (cock’s tail), 233 lava magma chamber
momentum-driven, 155, 159, 174, channel, 85, 389, 394 rupture, 14, 16
268, 370 delta, 275, 282 stresses, 14–16
surtseyan, 233 dome, 74, 129, 131, 145, 203, 339 fragmentation. See€fragmentation
vulcanian, 130, 141, 143 collapse, 368 mingling, 121
jetting, 368, 370, 375 fountain, 74, 153, 155, 168, 397 mixing, 18, 21–22
400
310
8/2/2012 12:27:14 PM
Figure 5.7╇╇ FLOWGO simulation of Mauna Loa’s 1984 flow 1A channel. Model run uses the input conditions for Mauna Loa’s
1984 flow given in Table 5 of Harris and Rowland (2001). In executing the model the software first projects a maximum gradient
path (in yellow) from the selected vent point to the edge of the DEM. The cooling-limited distance that the channel-contained
control volume can travel down that flow path is then projected in red. Comparison of model output and field measured
parameters are given in the inset box.
Figure 7.8╇╇ (a) Recent eruption of Fuego volcano (Guatemala) showing release of a buoyant eruption cloud (thermal). (b)
Streak lines illustrating velocity vectors for a laboratory experiment of an isolated thermal, showing upflow in center of thermal
and downflow at the edges (modified from Turner, 1973).
8/2/2012 12:27:38 PM
Figure 11.11╇╇ Explosive interaction after injection of water
into molten basaltic rock (Physikalisch-Vulkanologisches
Labor, Universität Würzburg).
0
1 km
(b)
2:00 hr
flow velocity
–1
(m s )
7
1 km
6
5
4
3
2
8/2/2012 12:27:43 PM
Figure 15.10╇╇ (a) Geometry of the upper 1 km of
conduit underlying the northern vent area of Stromboli. A
semi-transparent perspective cut-away view of the northwest
quadrant of the edifice including the sector graben of the
Sciara Del Fuoco, provides the reference for the location
and geometry of the conduit, which is derived from the
seismic source mechanisms obtained from inversions of
VLP signals associated with explosions. Surface illumination
from an external light source provides color contrasts,
emphasizing topographical features. A black line indicates sea
level. The summit of the volcano is 924 m above sea level (no
vertical exaggeration). The two flow disruption sites that are
sources of VLP elastic radiation are indicated by small white
circles. The irregular red and blue lines, respectively, mark
the surface traces of the dominant and subdominant dike
segments constituting the shallowest portions of the conduit
system. The eruptive vent is marked by a vertical arrow,
and temporary vents active during the flank eruption in
2002–2003 are indicated by green squares. The lateral extents
of individual dike segments are unknown and are shown
for illustrative purpose only. (b) Same as (a) for conduit
underlying the southern vent area. A slanted arrow points to
the southern vent. (c) Plunging view of the Sciara Del Fuoco
showing the two dike systems underlying the summit crater.
The conduit structure underlying the northern vent (marked
by a vertical arrow) is shaded green and that underlying the
southern vent (marked by a slanted arrow) is colored blue
(see (a) for details).
Figure 16.5╇╇ (a) Waveforms, (b) spectrogram, and (c) radiated acoustic power (in black) and ash cloud top (in green)
for the suplinian to plinian Tungurahua eruption of 17 August 2006. The eruption was characterized by continuous
jetting for >10 hours, with some explosions as well. Acoustic power broadly scales with ash cloud height for this
eruption, particularly during the plinian phase of the eruption around 17 August 0600 UTC. From Fee et al. (2010b).
8/2/2012 12:27:57 PM
Figure 17.6╇╇Volcanic features on Io. (a) Map of the Ionian surface showing key volcanic features. Image NASA/JPL/USGS;
merged Voyager and Galileo mosaic available at http://astrogeology.usgs.gov/Projects/JupiterSatellites/io.html. (b) These two
Galileo images show the appearance of a new, dark deposit 400 km in diameter that erupted from Pillan Patera in 1997, and
partly covering the large red deposit from Pele Patera. Image PIA00744, NASA/JPL/University of Arizona. (c) The New Horizons
Long Range Reconnaissance Imager captured this view of three of Io’s plumes on 28 February 2007. In the far north is the
390 km high Pele-type plume from Tvashtar volcano. In addition, a 60-km-high plume from Prometheus can be seen on the
western limb, and the top of Masubi’s plume can be seen poking into the sunlight from Io’s night side, to the south of the image.
Image PIA09248, NASA/Johns Hopkins Applied Physics Laboratory/Southwest Research Institute. (d) The floor of Tupan Patera
(75 km across) is partly covered with dark material, presumably lava, that appears significantly warmer in Near Infrared Mapping
Spectrometer data than the adjacent orange-colored “island,” parts of which are cool enough for SO2 to condense. Image
PIA03601, NASA/JPL. (e) The source of the Prometheus plume lies at the distal (westernmost) end of a compound pahoehoe
flow field. Image PIA02565, NASA/JPL/University of Arizona. Inset box shows the location of (f), which is a high-resolution
view of the active flow front. Dark areas on the flow show recent lava breakouts, older flow surfaces brghten with time due to
deposition of SO2 front from the plume. Fresh SO2 frost is also seen on the surrounding terrain. Image PIA02557, NASA/JPL/
University of Arizona. (g) Bright lobate features to the east of Emakong Patera have been interpreted as fresh sulfur flows. Image
PIA02539, NASA/JPL/University of Arizona.