Thouret 1999 Volcanic Geomorphology-An Overview

Ž .Earth-Science Reviews 47 1999 95–131www.elsevier.comrlocaterearscirev

Volcanic geomorphology—an overview

J.-C. Thouret )

( )IRD— Institut de Recherche pour le DeÕeloppement UR 6 and Centre de Recherches Volcanologiques, UMR 6524 CNRS Magmas et´Volcans, UniÕersite Blaise Pascal, Clermont-Ferrand, France´

Received 13 August 1998; accepted 17 February 1999

Abstract

The review examines the role of geomorphology in analyzing the volcanoes on Earth. Five objectives are stressed. First,classifications of volcanic landforms should be improved to take care of the complexity in volcanic landform generation asmagmatic systems, style of eruption, and the erupted material all influence the morphology. Second, geomorphology shouldcontribute to the science of volcanology through its capability in reconstructing growth ‘stages’ in complex volcanoes, andalso in analyzing the structural factors which contribute to the catastrophic collapse of volcanoes. Third, geomorphology cancontribute to physical volcanology by assessing the effects of topography on transport, erosion, and deposition ofvolcanogenic flows and identifying the sources and climaticrtectonic conditions which govern the emplacement of

Ž . Ž .volcaniclastic deposits. Fourth, volcanic geomorphology a identifies sedimentary facies associations, b constructs faciesŽ .models for dynamic volcano delivery systems, and c analyzes the characteristics of sediment gravity flows in order to

determine relevant parameters for modelling their behaviour. Fifth, process-oriented geomorphology is critical in developingaccurate methods for measuring rates of geomorphic processes that shape ephemeral volcanic constructs, and for evaluatingand comparing geomorphic impacts on disturbed catchments and the related hydrologic response before, during, and aftereruptions. This should help to refine parameters for the exponential decay model. Finally, volcanic geomorphology isessential for risk assessment through geomorphic hazard zonation and composite risk zonation. Such treatments arenecessary in order to face the enhanced challenge posed by the combination of natural hazards and the increasing number ofpeople who are at risk around volcanoes. q 1999 Elsevier Science B.V. All rights reserved.

Keywords: volcano growth; landforms; tectonics; debris avalanches; volcaniclastic sediments; denudation rate

1. Introduction

Volcanism directly creates and degrades land-forms, and indirectly provides an age for both thelandsurface over which the erupted material lies and

) Instituto Geofısico del Peru, Calle Calatrava 216, Urb. Camino´ ´Real, La Molina, Lima 12, Peru. Fax: q51-14-368-437; E-mail:[email protected]

the succession in which it is intercalated. Volcaniclandforms, in contrast to other types of landforms,result from both constructiÕe and destructiÕe forcesnear-simultaneously. Hence, volcanic landforms needto be studied carefully regarding their processes ofgrowth and erosion. In contrast to erosional land-forms in rocks surviving for long periods of time,volcanoes usually have a short-term existence, but

Ž .volcanic geomorphology enables us to 1 recon-

0012-8252r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0012-8252 99 00014-8

( )J.-C. ThouretrEarth-Science ReÕiews 47 1999 95–13196

struct original forms even after significant erosionŽ .has occurred, 2 estimate growth and denudation

rates based on geochronology and reconstructionŽ .techniques, and 3 compare the effects of climate on

the nature and rate of denudation. The last one isuseful for understanding similar volcanic landformsand rocks found over a wide range of climates.

We present a series of research themes in volcanicgeomorphology which stem from a review of papersrecently published both in volcanology and geomor-phology. Knowledge of volcanology and tools suchas DEM, GIS, and airborne imagery have expandedat an unprecedented rate over the past 20 years. We

Ž .intend to: 1 reappraise, on the basis of new toolsand methods, the terrestrial and marine landforms

Ž .generated by volcanic activity, 2 analyze the inter-play of construction and denudation processesthroughout eruptive activity in a variety of structural

Ž .and climatic settings, and 3 measure erosion pro-cesses acting on volcanoes and response of water-sheds disturbed by substantial eruptions.

2. Significance of volcanic geomorphology

The significance of Volcanic Geomorphology canŽ .be amplified through a the improvement of the

quantitative classification of volcanic landforms,which blends morphometry and studies based onground observations, remote sensing data, and labo-

Ž .ratory experiments, and b the diversified use ofairborne images and digital data acquired throughradar and satellites, and combined with DEM’s data,to facilitate the morphological analysis of volcanoes.

2.1. ImproÕement of classification of Õolcanoes andrelated landforms

Classical classifications of volcanic landforms arebased on types of activity, magmas, and erupted

Ž .products e.g., Cotton, 1944; Macdonald, 1972 . Im-proved classifications should also be based on geo-morphic scale, constructional vs. erosional origin,mono- vs. polygenesis, types of activity, and typeand volume of magma and erupted material. Here-after, we distinguish six main types of volcanic

Žconstructs and erosional landforms e.g., Ollier, 1988;

.Francis, 1993 . However, many rapidly constructedvolcanic landforms are not volcanoes at all, such asflood basalt, continental or submarine plateaus, andignimbrite sheets from large calderas.

2.1.1. Classification of Õolcanoes and related land-formsŽ .a Monogenetic landforms and fields:

Ž .-cinder or scoria cones, Surtseyan tuff cones, andŽ .Taalian tuff rings,

Ž .-maars subaqueous and subaerial and diatremes,Ž-intra- or subglacial volcanoes: tuyas table moun-

.tains and mobergs,-endogenous and exogenous domes,-lava flows and fields, including small-scale lava-flow forms,-continental flood basalts and plains basaltprovinces,-ash flows and ignimbrite sheets, plains, andplateaus.Ž .b Polygenetic Õolcanoes and calderas:-stratovolcanoes: simple with summit crater, com-posite with sector collapse scar andror a caldera;compound or multiple volcanoes,-intermediate-silicic multivent centres that lack acentral cone; rhyolithic centres; silicic volcaniclava field with multiple domes and calderas,

Ž .-calderas types: explosion somma , collapse-ex-Ž .plosion Krakatoa , collapse on Hawaiian shield

volcano, collapse in basement and resurgentŽ .caldera Valles , large and complex resurgentŽ .calderas Toba ,

-volcano-tectonic depressions termed ‘inverse vol-Ž .canoes’ Taupo Volcanic Zone .

Ž .c Shield Õolcanoes:-Hawaiian shields and domes; Galapagos, Ice-landic, and scutulum-type shields.Ž .d Volcanic landforms resulting from eruptiÕe

andror erosional processes:-avalanche caldera from a flank failure of mag-matic, gravitational, or mixed origin,

Ž-erosional calderas e.g., Haleakala, Maui; La.Reunion cirques .´

Ž .e Volcanic landforms resulting from denudationand inÕersion of relief:

-eroded cone; eroded pyroclastic-flow deposit andsheet,

( )J.-C. ThouretrEarth-Science ReÕiews 47 1999 95–131 97

-inverted small-scale forms: necks, culots, dykes,-eroded lava flow, inverted relief and planeze`Ž .e.g., Pain, 1995; Ollier, 1995 ,-roots of paleo-volcano, cauldron, and hypovol-canic complex.Ž .f Morphological changes in Õolcanic-surround

ing landscapes:-volcano construct and induced change in drainagepattern at a regional scale-drainage blockage, avulsion, impoundment andlake-breakout, etc.

2.2. Why do we need to improÕe existing classifica-tions?

2.2.1. SeÕeral examples illustrate the pitfalls of clas-sical classifications

The classification of tuff cones and tuff rings, forexample, has been based on effects of explosive

Žmagma–water interactions on morphology e.g., Cas.and Wright, 1987 : tuff rings and tuff cones are

thought to result from relatively dry and wet erup-tions, respectively, which are related to low and high

Ž . Ž . Ž .Fig. 1. After Sohn, 1996 his fig. 4, modified, caption modified . A Diagram of water–magma WrM ratio vs. explosion energy. Fourfields of hydrovolcanic eruption styles are arranged in a line along the X-axis, according to the assumption that the WrM ratio is the sole

Ž .controlling factor of hydrovolcanism. This model turns out to be flawed. B An alternative model in which each field of eruption stylesoccupies varying parts of the hydrovolcanic field depending on fundamental controls. According to this model, different styles of eruptionwith the same WrM ratio are possible, and an eruption may show only Taalian or Surtseyan characteristics even when it involves the fullgamut of high to low WrM ratios. Evolution paths of each tuff ring and cone, represented by thick arrows, are superimposed on the

Ž .diagrams. Reproduced with the kind permission to use material copyrighted by the Geological Society of America number 21569 .


Ž .mixing ratios of water to magma. Sohn 1996 haschallenged this prevailing model for tuff rings and

Ž .tuff cones in South Korea Fig. 1 , arguing that themorphological variations are directly caused by de-

Žpositional processes pyroclastic-surge dominated in.tuff rings and fallout-dominated in tuff cones , irre-

spective of water–magma mixing ratios. The deposi-tional processes are interpreted to be in turn con-trolled by a number of fundamental controls, whichinclude depositional settings, type, level, and lithol-ogy of aquifers, strength of country rocks, ground-water behavior, and properties and behavior ofmagma. These controls determine the explosiondepth, conduit geometry, mode of magma–waterinteraction, magnitude of explosion, eruption-columnbehavior, and subsequent depositional processes.

2.2.2. Existing classifications face increasing com-plexity in landform generation

Several examples illustrate the coexistence of sev-eral eruption styles and eruptive sequences withinthe lifetime of a complex volcano, and the contribu-tion of several mechanisms to caldera formation. Forinstance, giant tuff cones and calderas were de-scribed in arc volcanoes, such as the Ambrym caldera

Ž .and tuff cone, Vanuatu Robin et al., 1993 . Previ-ously considered as an effusive basaltic volcano,Ambrym consists of a basal shield volcano topped

by an exceptionally large tuff cone surrounding a12-km-wide summit caldera. The tuff cone may beconsidered as mainly basaltic. Interpretation of thetuff series implies intervention of external water andsuggests both explosive and collapse mechanisms forthis non-classical type of caldera formation at abasaltic volcano.

In addition, we need to improve the classificationto understand magmatic systems and achieve hazardassessments of the activity of a young volcano. Oneexample stems from the study of Cerro Negro vol-

Ž .cano, Nicaragua: McKnight and Williams 1997Ž .Fig. 2, Table 1 indicate that criteria of age and sizethat are usually used for this assessment are notadequate for active, young volcanoes. This questionbears on the fact that hazard assessment of thecurrent activity of Cerro Negro depends heavily onthe morphological type of volcano. Other criteria onwhich that determination can be based are magmaproduction rates, cone morphology, and eruptionstyle.

2.3. How can we improÕe existing classifications?

Detailed research has been undertaken over therecent years on volcanic landforms and on the mor-phology of lava flows, in terms of morphometry,comparative morphology, and processes, based on

Ž . Ž . Ž .Fig. 2. From McKnight and Williams, 1997 their figs. 4 and 5 . A Cumulative volume over age of volcano. Eruptivity rates er areŽ .shown. Solid line represents period 1850–1919; dashed line depicts time since 1923. B Crater and cone widths of Cerro Negro, determined

from historical accounts and photographs. Dashed line is trend for cinder cones; solid line is trend for composite volcanoes.


Table 1Ž .Summary of comparisons of cinder cones and composite volcanoes from McKnight and Williams, 1997, their table 2

Volcano Eruptivity Crater widthr Eruption Total Pyrocl. vol.r Cone Eruption Column3 3Ž . Ž .km rka cone width style vol. km total vol. height rate height

Ž . Ž . Ž . Ž .ratio DRE % km kgrs km1 4Paricutin 0.12 0.31 St 2.1 67 0.42 10 –10 0.1–6.0

Jorullo 0.12 0.25 St, V ;2.0 ;60 0.35 n.d. n.d.3 5Cerro Negro 1.6 0.08–0.36 St, V, Sp 0.22 77 0.25 10 –10 0.1–8.03aIzalco 8.8 0.07 St, V 2.1 ;18 0.65 10 FPlinian4 5bArenal 1.3 0.06 St, V, Sp, P 5.0 ;20 1.10 10 –10 FPlinian1 6Strombolian n.d. n.d. St – 1–20 – 10 –10 0.1–5

StsStrombolian, Vsvulcanian, Spssub-Plinian, PsPlinian, n.d.sno data, DREsdense rock equivalent.Ž . Ž . Ž .Data references: Paricutin—Foshag and Gonzalez 1956 , Fries 1953 ; Jorullo—Luhr and Carmichael 1985 ; Cerro Negro—McKnight

Ž . Ž . Ž . Ž . Ž . Ž .1995 ; Izalco—Rose and Stoiber 1969 , Mooser et al. 1956 ; Arenal—Wadge 1983 , Borgia et al. 1988 ; Strombolian—Wood 1980 .Ž . Ž .Production rate data from Hasenaka and Carmichael 1985 and Wadge 1982 .

a Ž .Calculated from Rose and Stoiber 1969 for the 1966 eruption only.b Ž .Calculated from Wadge 1983 for the 1968–1980 eruption only.

remote sensing, ground-based observations, and lab-oratory experiments.

2.3.1. Recent results acquired on large-scale land-forms

Shield volcanoes, for example, have been thefocus of recent morphological studies on size, distri-

Žbution, and magma output rate Mexican shield vol-.canoes: Hasenaka, 1994 , and on morphology andŽmechanism of eruptions Icelandic shield volcanoes:

.Rossi, 1996 . Additional results, acquired on smallerlandforms, such as dome origin and behaviour, illus-trate the evolution from a descriptive approachŽ .Scarth, 1994 to a semi-quantitative approachŽ .Blake, 1990 .

2.3.2. Detailed morphology of laÕa flowsRecent studies on lava flows have been based on

three approaches: fieldwork and morphometry, re-mote sensing, and laboratory experiments.

Ž .a Ground observations and quantitative, compar-ative morphology have beeen carried out on ice-landic lava flows and small-scale landforms on lava

Ž .flows Rossi and Gudmundsson, 1996 .Ž .b Remote sensing helps to outline lava flows

and detail lava-flow morphology using spaceborneŽ .Radar images TOPSAR, SIR-C radar , satellite im-

Ž .ages Landsat TM, SPOT , and photographic data.Digital multispectral data such as the thermal in-

Ž .frared multispectral scanner TIMS images help to

infer the chemical and physical properties of thesurface materials and to map lava flows. For exam-ple, the Mauna Loa lava flows were mapped in great

Ž .detail Kahle et al., 1995; Kauahikau et al., 1995using NASA’S TIMS and Space Shuttle radar SIR.Lava flows can be followed up and mapped usingdata obtained from the spaceborne advanced very

Ž .high resolution radiometer AVHRR , whose quanti-tative analysis allows estimation of active lava area,thermal flux, effusion rates, and total flow fieldvolume. Estimates of eruption rate and total flowfield volume of the 1991–1993 Mount Etna effusive

Ž .eruption Harris et al., 1997 are in agreement withpublished ground-based estimates of 5.8 m3 sy1 and

6 3 Ž235=10 m Calvari et al., 1994; Tanguy et al.,. Ž .1996 . Stevens et al. 1997 carried out a more

accurate estimate of the 1991–1993 lava-flow vol-Ž 6 3.ume 231"29=10 m , using EDM-based field

survey of the surface of the lava-flow field and onepanchromatic SPOT image. The results were digi-tised, interpolated and converted into a DEM, con-structed from a 1:25,000 contour map of the area.Digital elevation data from TOPSAR, an airbornesynthetic aperture radar system that uses interferome-try to derive topography, were used by RowlandŽ . Ž .1996 Fig. 3A,B to determine slope distributions,proportions of lava flows and vents, and lava flowthicknesses and volumes on Fernandina Volcano,Galapagos Islands. The concentration of vents on thesummit platform, five to eight times as much as on


Ž . Ž .Fig. 3. After Rowland, 1996, his figs. 7 and 9, modified, caption modified . A Vent map superimposed on a shaded relief image. Ventsare mapped from SPOT and TOPSAR data. An attempt was made to draw the symbol at each vent location to match the extent of the actualpyroclastic construct, but for some of the narrower arcuate fissures the pyroclastic deposit is less distinct than the mapped symbol. Note thestrong differentiation of vents into arcuate and radial categories. Note also that the distribution of radial vents is not uniform. Dotted circleŽ . Ž .centered on the caldera includes 95% of all radial vents and has a radius of 13 km. B Graph of young vent elevation versus distance fromthe center of the volcano, differentiating vents by their orientation, volume of lava produced, and type of lava produced. Note that radialvents have produced the largest volume of lava, and that pahoehoe has been preferentially produced at, or within 3 km of, the base of thesteep slopes. This graph approximates a radially averaged topographic profile with a vertical exaggeration of ;4= . There are fewer circlesthan young flows because some vents have been buried and because there are young flows for which vents cannot be identified. Reproduced

Ž Ž . .with the kind permission of American Geological Union Journal of Geophysical Research, 1996, 101 B12 , 27657–27672 .


coastal plain and apron, supports previously pro-posed mechanisms for producing higher elevationsand steeper slopes in the central part of the volcano.

Ž .c Laboratory simulations and comparative mor-phology have been carried out to reproduce mor-phologies observed on the sea floor. Using labora-

Ž .tory simulations, Gregg and Fink 1995 were able toreproduce submarine flow morphologies commonlyobserved on the sea floor, typically classified aspillowed, lobate, or sheet flows, while monitoringthe physical conditions under which they form. Foursubmarine lava-flow morphologies are considered tobe diagnostic of specific effusion rates: jumbled,folded, and lineated sheets, and striated pillows.

2.3.3. A wealth of Õolcanic landforms on oceanfloors

Recent investigations confirm that the oceanfloors, in particular the mid-ocean ridges, are hometo over 60% of the Earth’s volcanoes. Investigationsof oceanic ridges have emphasized the fast-spreadingEast Pacific Ridge, the medium-spreading Juan deFuca Ridge, the slow-spreading Mid-Atlantic Ridge,and the super slow-spreading SW Indian Ridge. Vol-canic constructs include axial topographic highs,abyssal hills, and seamount populations which showa spatial density and characteristic height in accor-

Ždance with the spreading rate Smith and Cann,.1992; Mendel and Sauter, 1997 .

Ž .a Based on swath bathymetric coverage com-bined with high-resolution side-scan images, thethree-dimensional perspective view of the axis of the

Žslow-spreading Mid-Atlantic Ridge Smith et al.,.1997 , shows volcanic constructs and faults similar

in size and shape to those observed at subaerial riftzones such as Hawaı and Iceland. The overall shapeöf the axial zone is that of a major graben composedof an inner valley floor and bordered by valley wallsalong normal faults. The inner valley floor is theprimary site of crustal construction, and most seg-ments contain large axial volcanic ridges within theirvalley floors that are the principal sites of lavaextrusion: seamounts, hummocks, fissures, andsmooth flows. Axial volcanic ridges range in size upto several hundreds of meters high, 1.5 km wide, and

Žseveral to tens of kilometers long. Small 50-h-.300 m near-circular seamounts are distributed over

the valley floors.

Ž .b The topographic features known as abyssalŽhills typically 10–20 km long, 2–5 km wide, 50–300

m high, and oriented approximately perpendicular to.the spreading direction , characterize )30% of the

Ž .ocean floor Macdonald et al., 1996 , being the mostabundant geomorphic structures on Earth. Sub-mersible-based investigations show that Pacificabyssal hills are created on the East Pacific Rise ashorsts and grabens which lengthen with time. Hillsare bounded on one side by ridge-facing scarpsproduced by normal faulting, and on the other bymore gentle slopes produced by volcanic growthfaulting.

Ž .Seamounts Carlowicz, 1996; Smith et al., 1997 ,Ž .guyots Smoot and King, 1993; Smoot, 1995 , and

Ž .shoaling volcanoes McPhie, 1995 play a significantrole in crustal construction and in constructional-ero-sional processes. Seamounts play a significant role incrustal construction at the mid-oceanic ridges, atleast in the slow-spreading ridges such as MAR.Spreading segments contain a prominent axial vol-canic ridge. Ridges are composed of piled upseamounts and hummocky flows, and are interpretedas the primary sites of crustal construction. Smallmagma pockets with slow eruption rate produceseamounts; small magma bodies with somewhathigher eruption rates produce hummocky fissure fedflows.

2.4. Geodynamic and tectonic settings of Õolcanicconstructs

The passage of magma through the crust andlithosphere is controlled by crustal lithosphericstress-field and local stress-field configurations, andresultant fractures, i.e., normal faulting, thrust fault-

Ž .ing, and strike–slip faulting Cas and Wright, 1987 .

2.4.1. Tectonic effects on Õolcano and caldera loca-tion, morphology, and formation

To show the relationships between volcanic com-plexes or caldera location, morphology, and tecton-

Ž .ics, three approaches have been undertaken 1 on aŽ .morphotectonic basis, 2 on structural and remote

Žsensing data e.g., shape and formation of large. Ž .calderas , and 3 more recently, on laboratory exper-

iments and modelling.Classical morphotectonic studies include the sta-

tistical analysis of the geometry of drainage patterns


and stream directions controlled by tectonics and theidentification of morphotectonic features. For exam-ple, the main morphological characters of volcanic

Žcomplexes of Latium in Italy Buonasorte et al.,.1991; Trigila, 1995 are strongly controlled by four

prevailing tectonic directions for Vulsini, and threefor the Sabatini and Colli Albani areas. These studiesallow us to infer where and how the structural settingof the sedimentary units and the recent tectonicactivity of the area ‘control’ the location and shapeof the calderas.

The influence of the structural framework on theshape and the formation of large caldera complexeshas been inferred from remote sensing and ground-based structural analysis. Based on 2D-images pro-duced by draping SPOT satellite images over aDEM, the relationships between the geodynamic set-ting, the regional faults, and the calderas of Toba and

Ž .Tondano Indonesia are described in terms of evolu-Ž .tion in a pull-apart basin Lecuyer et al., 1997 . The

links between the main morphologic and structuralfeatures on the submerged portions of volcanic edi-fices such as Panarea, Stromboli in the Aeolian

Ž .Islands Gabbianelli et al., 1993 , have been carriedout through electroacoustic and high resolution seis-mic profiles in the SE Tyrrhenian sea. Both com-plexes show a preferential development along NE–SW lineaments, which coincides with the regionalstructural trend of this sector of the Aeolian struc-ture. Faulting, caldera collapse, and tectonic tiltingwere interrelated and fundamentally influenced byactivation of the NE–SW fractures.

2.4.2. Volcanic constructs can influence their tec-tonic setting and magmatic system

Ž .a Volcano building on thin oceanic lithospherecauses the lithosphere to sag downward into the

ŽFig. 4. From Van Wyk de Vries and Merle, 1996 their fig. 1,.caption modified . Examples of influence of volcanoes on fault-

Ž . Ž .ing. A Fieale volcano, showing inward-curving faults. B Axialand Brown Bear seamounts, showing Juan de Fuca Ridge curving

Ž .into volcanoes. BBB, Brown Bear basin; HB, Helium basin. CMaderas volcano. Thick lines indicate major faults. This volcanospreads and rifts predominantly normal to regional extensionŽ .indicated by arrows . Shaded areas indicate upper part of eachvolcano. Contours are in metres. Reproduced with the kind per-

Žmission of the Geological Society of America GSA Copyright.permission number 21569 .

asthenosphere. As much as a half or two-thirds of thebuild-up of the volcanoes of Hawaı may be offset by¨


Žlithospheric subsidence Peterson and Moore, 1987;.Lipman, 1995 .

Ž .b The tectonic effect of volcanic constructs onfaulting has been tested through laboratory experi-ments and modelling. The effect of volcanic con-

Žstruct on rift fault patterns Van Wyk de Vries and.Merle, 1996; Fig. 4 is exemplified in three cases:

Ž .Fieale volcano Asal Rift, Djibouti , Axial and BrownŽ .Bear seamounts Juan de Fuca Ridge , and Maderas

Ž .volcano Nicaragua . Analogue models show thatincreased fault throw as the volcano is approached iscaused by an interaction of the regional stress fieldwith that set up by the volcano mass. For faults to bereoriented there must be a ductile layer below the

Žvolcano hot crust at mid-ocean ridge, weak sedi-.mentary strata, etc. . Increased volcano mass and

size and lower brittlerductile ratios lead to increasedfault curvature. Volcanoes on one side of a rift maycapture the fault, forming the axis of a new rift. Byconcentrating extension, magma is more easilyerupted. A positive feedback between increased ex-tension and magma eruption rate will lead to riftnarrowing, which can favor the formation of oceanic

Ž .crust Van Wyk de Vries and Merle, 1996 .Ž .c Conversely, volcanic constructs and gradual

volcano spreading influence faulting and riftingŽMerle and Borgia, 1996; Van Wyk de Vries and

.Merle, 1996 , as well as the slope instability of theŽ .edifice Van Wyk de Vries and Francis, 1997 . A

volcano of sufficient size induces stresses that maydeform its substratum. In turn, this deformation feedsback stresses which deform the edifice. Both stressesand deformation influence the evolution of magmaby varying the boundary conditions of magmaticsystems.

3. Contribution to geology: volcano growth anddestruction, large-scale instability, and relation-ships with tectonics and sedimentation

To contribute to geology, volcanic geomorphol-Ž .ogy should a produce detailed geomorphological

maps and expand the use of accurate chronologicalframeworks and compositional data through eruptivesequences, to identify eruptive or constructional

Ž .‘stages’ in complex volcanoes, b expand the capa-bility of geomorphology in landscape history recon-

struction, thus contributing to the understanding ofprocesses of building and destruction of volcanic

Ž .edifices, and c analyze the structural factors whichcontribute to the catastrophic collapse of volcanoes.

3.1. Growth and destruction–denudation of Õolca-noes

Growth and destruction of volcanoes are the resultof the complex interplay of endogenous and exoge-nous processes. One of the major tasks of volcanicgeomorphology is to reconstruct the volcanic land-

Ž .scape and landform history, in order to 1 unravelthe building and destruction stages and processes,

Ž .and 2 calculate long- and short-term growth anddenudation rates over time, in response to eruptivestyle, tectonic uplift or spreading, climate, and sea-level change.

3.1.1. Rapid processes and rates of growth anddestruction of Õolcanic constructs

Unlike ordinary mountains, which are formed byslow uplift and erosion, volcanoes are constructedrapidly.

3.1.1.1. Growth and mature stages of oceanic shieldÕolcanoes. Seven stages characterize the evolution ofthe oceanic shield volcanoes located above an active

Žhot spot, such as the Hawaıan volcanoes Petersonänd Moore, 1987; Decker et al., 1987; Moore andClague, 1992; Tilling and Dvorak, 1993; Rhodes and

. Ž . Ž .Lockwood, 1995 : 1 initial stage; 2 shield-build-ing stage including three submarine, sea-level, and

Ž . Ž .subaerial substages; 3 capping stage; 4 erosionalŽ . Ž .stage; 5 renewed volcanism stage; 6 atoll stage;

Ž .and 7 late seamount stage. The seven volcanoescomprising the island of Hawaı and its submarine¨base are, in order of growth, Mahukona, Kohala,Mauna Kea, Hualalai, Mauna Loa, Kilauea, and thestill submarine volcano, Loihi. The first four havecompleted their shield-building stage.

The island of Hawaı has grown at an average rateöf about 0.02 km2ryr for the past 600 kyr and

Žpresently is close to its maximum size Moore and.Clague, 1992; Fig. 5A . On each volcano, the transi-

tion from eruption of tholeiitic to alkalic lava occursnear the end of shield building. The rate of south-eastern progression of the end of shield building, and


hence the postulated movement rate of the Pacificplate over the Hawaıan hotspot, in the interval from¨Haleakala to Hualalai is about 13 cmryr. Based onthis rate and an average spacing of volcanoes of40–60 km, the volcano requires about 600 thousand

years to grow from the ocean floor to the time of theend of shield building. They reach the ocean surface

Žabout midway through this period Moore andClague, 1992; Lipman, in Rhodes and Lockwood,

.1995; Fig. 5B .


3.1.1.2. Caution needed to reconstruct the long-termeÕolution of stratoÕolcanoes. Recent detailed studiesof time–volume-composition concerning the long-

Žterm behavior of stratovolcanoes in active arcs e.g.,Mount Adams, Cascades: Hildreth and Lanphere,1994; Tatara–San Pedro complex, Chile: Singer et

.al., 1997 challenge previous studies in stating thatcaution should be applied in reconstruction of thelife history of so-called polygenetic Õolcanoes.

Subdivision of complex stratovolcanoes into erup-tive and constructional ‘stages’ needs detailed geo-logic mapping, accurate high-resolution geochronol-ogy, and compositional data. Stratovolcanoes com-monly grow in spurts: construction of an imposingcone needs take only 1–5% of the active lifetime ofthe volcano, i.e., construction rate of 1–5 km3rky.

ŽHigher rates are exceptional Hildreth and Lanphere,.1994 . Stratovolcanoes can remain active between

the widely spaced episodes of peak productivity, asmuch as for half a million years. Documented exam-ples of greater longevity are rare. ‘Dormancy’ is ananthropocentric notion and generally refers only toan upper-crustal condition, without fundamental

Žchange in deep-level magmatic processes Hildreth.and Lanphere, 1994 .

Stratovolcanoes need not develop large upper-crustal magma chambers and need never evolve to-wards a caldera-forming stage. Arc calderas thatresult from collapse of shallow reservoirs beneath

Ž .stratovolcanoes Mazama, Krakatau, Santorini areusually associated with large eruptions of rhyodacitic

to rhyolitic ejecta. Recurrent eruptions of smallbatches of dacite at irregular intervals and its secularalternation with varied andesite and even andesiticbasalt shows that there is no standard sequence, nounidirectional progression, and certainly nothing pre-determined in the evolution of stratovolcanoesŽ .Hildreth and Lanphere, 1994 .

Andesite–dacite production in the focal regionand coeval basaltic activity on the periphery have

Žcoexisted at several documented stratocones e.g.,.Mount Adams: Hildreth and Lanphere, 1994 .

Scarcity or abundance of surrounding mafic cindercones and lavas has nothing to do with maturity ofthe stratovolcano system. The term parasitic shouldbe abandoned as it implies dependence upon themain stratocone and promotes the view that periph-eral mafic eruptions are leaks from a central chamberor conduit. Virtually the opposite is true: andesiticstratovolcanoes are the derivative features and flankfailures are lateral breakouts from a central conduitsystem, whereas peripheral basalts, having their ownconduit from mantle or deep-crustal depths, are more

Ž .fundamental Hildreth and Lanphere, 1994 .

3.1.2. Volcano growth and erosion result from com-plex and interdependent processes

3.1.2.1. Erosion-prone glacial periods haÕe alteredthe fast growth record of mid-latitude stratoÕolca-noes. The life history of Andean volcanoes providesan important insight into how we may interpret the

Ž . Ž .Fig. 5. A From Moore and Clague, 1992, their fig. 8, caption modified . Estimated ages for events in the life history of volcanoes on oradjacent to the island of Hawaı. Volcano positions are projected and measured on a N408W line passing through Kilauea. Squares indicate¨

Ž .timing of transition from eruption of tholeiitic to alkalic lava two symbols where two transitions were dated . Solid line is least-squaresŽ .regression line through points representing the end of shield building of five volcanoes solid circles ; its slope is 13 cmra, the apparent rate

Ž .of progression of volcanism in this segment. Upper dashed line birth is drawn parallel with lower one through the postulated position ofŽ . Ž .Keikikea large circle assuming zero age of birth; second dashed line end of early alkalic stage is drawn through position of Loihi at zero

Ž .time; third dashed line emergence is drawn midway between birth and end of shield building. Reproduced with the kind permission of theŽ . Ž . Ž .Geological Society of America GSA Copyright permission number 21569 . B From Lipman, 1995 his fig. 12, caption modified .

Ž . Ž .Interpreted growth history of Mauna Loa, showing inferred magma-supply rates solid line and cumulative volumes dashed line with time.Ž . Ž .Curves are constrained by: 1 present total volume and magma-supply rates, 2 requirement for higher magma-supply rates during peak of

Ž . Ž .tholeiitic shield-building stage, 3 lower rates during probable early alkalic stage at Mauna Loa and analogous to Loihi, 4 indication fromthe declining recent eruption rates and helium-isotope evidence that tholeiitic shield building is nearing completion and alkalic volcanism

Ž .should be expected at Mauna Loa in the not-too-distant geologic future, and 5 overall estimated lifespan of about 1 Ma for Hawaıan¨Žvolcanoes. Reproduced with the kind permission of American Geophysical Union Mauna Loa revealed: Structure, Composition, History,

.and Hazards, 1995 .



Ž .Fig. 6 continued .

chronological record of glacially eroded volcanicedifices. The Tatara–San Pedro complex in the

ŽNorthern Chilean Andes Singer et al., 1997; Fig.. 36A preserves about 55 km of lavas that erupted

during seven eruptive sequences from at least threecentral vent regions. Remnant, unconformity-boundsequences of lavas are separated by lacunae thatrepresent significant periods of erosion. Estimatedgrowth rates for the two young volcanoes are 0.2 to

3 Ž .0.3 km rky Fig. 6B , i.e., three to five times greaterthan a growth rate estimated from all preserved lavas

Ž 3 .in the complex 0.06 km rky . Removal of up to50–95% of the material erupted between 930 and

200 ka by repeated glacial advances largely explainthis discrepancy, and it raises the possibility thatepisodic erosion of mid-latitude frontal arc com-plexes may be extensive and common. Hence, Singer

Ž .et al. 1997 raise fundamental questions bearing onthe interpretation of the life history of glaciallyeroded stratovolcanoes.

Ž .Frontal arc volcanic complexes of the Andesmay remain active for about 1 my, during whichtime many changes in vent position, magma compo-sition, and magma reservoir processes can occur.Thus, several earlier volcanoes of possibly compara-ble dimensions can be largely obliterated. Preservedlavas and erosional hiatuses in dissected volcanoes,consistent with global records of terrestrial ice ad-vances, offer additional valuable source of informa-tion bearing on the timing and extent of glaciations,to bracketing the age and extent of tills, drifts, andmoraines in regions of high elevation or maritimeclimate.

Accumulation of ice volumes leading up to glacialmaxima is a slow process requiring 80–150 ky. By

Žcontrast, major deglaciations proceed rapidly -15.ky to glacial minima. In contrast to selectively

eroded young units and summit portions of anyvolcanic edifice, the oldest lava at the base of asequence overlying an erosional unconformity maybe emplaced very shortly after the glacial maximum.Thus, the correspondence in timing between lacunaein stratigraphic sequences and the global ice-volumemaxima recorded in the astronomical time scale can-

Žnot be considered merely fortuitous Singer et al.,.1997 .

Ž . Ž . 40 39Fig. 6. After Singer et al., 1997 their figs. 5 and 8, modified, caption modified . A Stratigraphy, K–Ar and Arr Ar ages, andpaleomagnetic orientations of the Tatara–San Pedro complex. The astronomical time scale is based on marine oxygen isotope data. Evennumbered peaks in E

18 O reflect global ice-volume maxima. Note that the oldest lavas in the Quebrada Turbia, Estero Molino lower, EsteroMolino upper, Guadal–Placeta San Pedro, and Volcan San Pedro units provide minimum ages for the unconformities that they overlie. Ages´

Ž . 18of these basal lavas correspond to periods of rapid deglaciation gray shading following E O peaks at stages 20, 16, 10, 8, and 2. Volcan´Ž . Ž .Tatara is younger than the penultimate glaciation stage 6, 130 ka , but older than the last glaciation stage 2, 17 ka . In addition, a gap in

the ages of preserved lavas of Volcan Pellado corresponds to the stage 6 maximum. Late activity at Volcan Pellado overlaps in age withŽ . Ž . 40 39basal Tatara lavas dashed horizontal line . Between three periods 930–770 ka, 605–220 ka, and 190–19 ka , the K–Ar and Arr Ar ages

are arranged in ascending stratigraphic order from left to right. Three 40Arr39Ar ages from Cordon Guadal are shown as open symbols. The´Ž .Matuyama–Brunhes M-B polarity reversal is 775"10 ka in the astronomical time scale, identical with the ages obtained from

Ž . Ž .transitionally magnetized Quebrada Turbia lavas N—normal, R—reversed, and T—transitional orientation . B Cumulative eruptivevolume vs. time for the Tatara–San Pedro volcanic complex. Inset illustrates models projecting the growth of the complex at three different

Žrates, assuming that no erosion took place. Reproduced with the permission of the Geological Society of America GSA Copyright.permission number 21569 .


3.1.2.2. RepetitiÕe and instantaneous mass-wastingdestruction characterize the deÕelopment of oceanicÕolcanoes. In the Canary islands with Quaternaryvolcanic activity, the construction has been closelycontrolled by rift-type volcano-tectonic features,characterized by a tight cluster of recent emissioncentres piled up along narrow dorsal ridges. Twomain types of rifts can be defined in terms of geome-try: simple, as in La Palma, or triple, as in Tenerife

Ž . Ž .and Hierro. Carracedo 1994 Fig. 7A provided ahotspot-based schematic model for the genesis of a

complex ‘Mercedes’-type stellate rift zone on one ofthe Canary Islands. The geometry of the complexCanary three-branched rifts separated by angles of

Ž .1208 Fig. 7B suggests a least-effort fracture as aresult of magma-induced vertical upwards loading.The concentration of the recent eruptive activity andthe depressions that may have been generated bygravitational slides are also included in the modelŽ .see discussion by Marti et al., 1996 . The impor-tance of this model is that it may help to explain themain landforms and the two sources of volcanic

Ž . Ž .Fig. 7. A From Carracedo, 1996a his fig. 7 . Contrasting evolutionary patterns and volcanic hazard sources in the western and easternislands of the Archipelago. The main factors controlling these differences are the activity of a hotspot and the lateral variation of the

Ž . Ž .structure, thickness and rigidity of the crust in the oceanic west end and the transitional eastern end, near the African coast sectors of theŽvolcanic chain. Reproduced with the kind permission of the Geological Society Volcano instability on the Earth and other Planets, special

. Ž . Ž .publication 110, pp. 125–135 . B After Carracedo, 1994 his fig. 6, modified . A hotspot-based schematic model for the genesis of acomplex «Mercedes»-type stellate rift zone, on one of the Canary Islands. The concentration of the recent eruptive activity and the

Ždepressions that may have been generated by gravitational slides are also indicated. Reprinted from Journal of Volcanology and GeothermalŽ . .Research, 60 3–4 , 225–241, Copyright 1994, with permission from Elsevier Science, 1999 .


Ž .Fig. 7 continued .

Ž .hazards on the Canary Islands Carracedo, 1996a,b :Ž . Ž1 the concentration of recent Quaternary–Holo-

.cene eruptive vents, and therefore, the statisticallyŽ .most probable location of future eruptions, and 2

the genesis of the main open-towards-the sea Canary

depressions and, subsequently, the areas where catas-trophic slope failures may be pending.

In the Canaries, two main types of large depres-sions are open toward the sea: straight-walled andarcuate head basins such as the Orotava and Guimar


valleys in Tenerife and crescent-shaped coastal em-bayments like El Golfo in Hierro island. Their gene-sis has been the subject of a long debate. Faulting

Žand collapse in giant gravitational landslides Car-.racedo, 1996a,b; Masson, 1996 are now preferred to

a mainly erosional origin. An alternative hypothesisof lateral expansion of the Guima and Orotava val-leys based on multiple cycles of erosion and fillingof valleys by lava flows has also been proposedŽ .Palacios, 1994 . However, recent detailed mappingof areas of the submarine flanks of the Canaries hasrevealed much new evidence of seven major land-slides over the past 500,000 years, strengthening thecase for large-scale slope failure as a principal agent

Ž .of island destruction Masson, 1996 .

3.2. Large-scale instability processes, erosionallandforms, and debris-aÕalanches

As a consequence of rapid construction, manyvolcanoes are liable to massive flank or slope fail-ures resulting from structural instability. More than20 major slope failures have occured globally duringthe past 500 years, a rate exceeding that of caldera

Ž .collapse Siebert, 1996 . Slope failures produce ex-tremely mobile debris avalanches that can travel longdistances beyond the flanks of volcanoes at highvelocities. The characteristics and origins of flankfailures and related deposits were described by Voight

Ž . Ž . Ž .et al. 1983 , Siebert et al. 1987 , Crandell 1988 ,Ž . Ž .Glicken 1991 , Moore et al. 1994 , McGuire et al.

Ž .1996 , as follows.Massive landslides create specific morphology and

deposits, i.e., horseshoe-shaped re-entrants into theedifice, and a high, steep-sided break-away scarphaving an amphitheatre shape. Debris avalanchestypically form a hummocky terrain with water-filleddepressions and steep flow margins, and thick hum-mocky deposits with block and matrix facies oflargely unsorted and unstratified angular-to-subangu-lar debris. A relationship exists between the distancerunout travelled by an avalanche and the failurevolume. The maximum failure volume of subaerialvolcanoes typically does not exceed 10% of theedifice volume. The ratio of vertical drop H to travel

Ž .length L range from 0.09 to 0.18 av., 0.13 forQuaternary volcanic avalanches between 0.1 and 1

3 Ž .km in volume and from 0.04 to 0.13 av., 0.09 for

3 Ž .avalanches )1 km Siebert et al., 1987 . The ratioof H to L for volcanic avalanches is much lowerthan the ratio for non-volcanic deposits of similarvolume, suggesting that low-rigidity, perhaps par-tially fluidized avalanches are capable of travellinggreat distances. Exceptional runout distance G100km travelled by avalanches have been reported atstratovolcanoes such as Nevado del Colima in Mex-

Ž .ico Stoopes and Sheridan, 1992 .The widespread occurrence of slope failure in a

variety of tectonic settings suggests that it may bethe dominant catastrophic edifice-modifying process.Steep-sided andesitic and dacitic stratovolcanoes,with a relief that can attain several km and upperslopes that can exceed 308, are obvious candidates

Žfor slope failure e.g., Socompa volcano, Northern.Chile: Wadge et al., 1995 . Steep-sided but less

voluminous lava-dome complexes are also particu-larly susceptible to slope failure. The air of perma-nence of large, low-angle shield volcanoes beliestheir inherent instability. Particularly noteworthy are70 landslides that have occurred on the Hawaiianridge, where they have removed volcano-flank sec-

3 Žtors that exceed 1000 km in volume Moore et al.,.1989, 1994; Iverson, 1995 . A series of maps from

several sources document the importance of land-slides, best exposed in the submarine realm, in thegrowth and decline of Mauna Loa and adjacent

Ž .volcanoes Moore and Chadwick, 1995 . Catas-trophic slope failures are neither rare nor unique inthe lifetime of a volcano. The summit edifice ofAugustine, AK, has repeatedly collapsed and regen-erated, averaging 150–200 years per cycle, during

Ž .the past 2000 years Beget and Kienle, 1992 . Theunprecedented frequency of summit edifice failurewas made possible by sustained lava effusion ratesover 10 times greater than is typical of plate-marginvolcanoes.

The origins of flank failure are bound to threeŽ .types of events Siebert et al., 1987; Siebert, 1996 :

Ž .magmatic eruption of Bezymianny type 1956 ,Ž .non-magmatic explosions of Bandai type 1988 , and

Ž .cold avalanches of Ontake type 1784 . Severalstructural and geomorphic factors contribute to flank

Žfailure Voight et al., 1983; Vallance et al., 1995;. Ž .McGuire et al., 1996 : 1 steep dip slopes with

alternating competent lavas and unconsolidated pyro-Ž .clastic materials; 2 zones of weakness within the


upper volcanic edifice, owing to hydrothermally al-tered rocks converted to clay minerals of low yieldstrength, accompanied by the boiling of supercritical

Ž .hydrothermal fluids Lopez and Williams, 1993 ; andŽ .3 local extension promoted by parallel dike swarmsŽ .Hausback and Swanson, 1990 .

Yet important but enigmatic questions remain un-solved. What processes can trigger slope instabilityon low-angle shield volcanoes? Forces in addition to

Žgravitation must trigger the landslides Iverson,. Ž .1995 . Tilling and Dvorak 1993 invoke a seaward

displacement of the southern flank of Kilauea vol-Ž .cano, and Owen et al. 1995 measured its rapid

deformation to be as much as 10 cm ay1, based onGPS measurements. The observations can be ex-plained by slip on a low-angle fault beneath thesouth flank combined with dilatation deep withinKilauea’s rift system, both at rates of 10–15 cm ay1.

What fundamental structural causes might lead toan edifice collapse? Van Wyk de Vries and FrancisŽ . Ž .1997 Fig. 8 , diverging from previous works whichemphasized differences in eruption style associated

Ž .with flank failure Siebert et al., 1987 , argue that

ŽFig. 8. After Van Wyk de Vries and Francis, 1997 their fig. 3,.modified, caption modified . Interpretation of spreading and col-

lapse structures at Mombacho, showing contrasting models forŽ . Ž .basement failure at Las Isletas a and b , El Crater c and

Ž .probable precursory features. a Las Isletas: basement collapse.Ž .Deep spreading on decollement X within the Las Sierras Forma-

tion rises to produce a frontal anticline at the foot of Mombacho.Spreading induces differential movements within the upper parts

Ž .of the volcano, contributing to the initiation of slip on a plane Ywithin the cone. Once the dip-slip decollement is activated, themass above it places additional load on the spreading front,inducing increased movement, and eventual failure through the

Ž . Ž .frontal anticline Z . b Plan diagrams illustrating the differencein structural style between a radially spreading and a sector

Ž .spreading volcano that is, Mombacho . Radial spreading pro-duces inward dipping normal faults that cut any potential failureplane in the cone. In contrast, sector spreading creates outward

Ž .dipping faults, which promote collapse. c The El Crater collapsecrater is ;1.5 km wide and long, and 700 m deep. Its walls curveinwards toward the opening, where there is a pronounced 30-m-

Ž .high lip L . Overall, the shape is that of a rotational slump failurein a mechanically homogenous medium, so no pre-existingdecollement place is required. Such failures usually begin whenshear strength is reduced over a wide area. At El Crater thishomogeneous strength loss was produced by progressive hy-drothermal alteration.

the volcanic edifice itself can contribute to the weak-ness of its bedrock. In contrast to radially spreadingvolcanoes, preferential spreading in one direction iscritical to collapse development; whereas radialspreading tends to generate inward-dipping faultswhich inhibit collapse, sector spreading generatesfailure-prone outward-dipping structures. Spreadingin a preferential direction may be caused by buttress-ing, by the regional slope of basement beds, byregional stress, by weak basement or by high fluid

Žpressures under one side Van Wyk de Vries and.Francis, 1997 .

l-32


3.3. Relationships between Õolcanism, tectonics, sed-imentation, and drainage

Volcano growth and erosion, as well as tectonics,influence sedimentation and the evolution of drainagepatterns in volcanic landscapes.

3.3.1. Modification of drainage patternIn his study of the morphotectonic development

Ž .of southeast Australia, Ollier 1995 shows severalexamples of drainage modifications in catchmentsaffected by the formation and erosion of huge volca-noes and by the retreat of the Great Escarpement.Large volcanoes on or near the Great Divide resultedin the formation and superimposition of radialdrainage in the vicinity of the volcano, and majordrainage disruption in neighbouring regions, evenaffecting large rivers.

3.3.2. Complex relationships between Õolcanism, tec-tonics, and sedimentation

The geomorphic evolution of the Sunda volcaniccomplex and the Bandung area, a large intramontane

Žbasin surrounded by volcanic highlands in Java Dam.et al., 1996; Nossin et al., 1996 , illustrates complex

landform-determining processes, such as tectonicsubsidence, paroxysmal eruptions, volcanism-in-duced faultingrrifting, drainage system adaptations,and intramontane lacustrine sedimentation. The mor-phology of the Bandung basin and the Sunda–Tangkuban Perahu volcanic complex encompassedseven phases during the Middle–Late Quaternary, in

Ž .particular since 125,000 yr B.P.: 1 the early Ban-Ž .dung basin during Middle Quaternary; 2 the start of

lacustrine sedimentation and the formation of anŽ .enclosed intramontane basin ca. 125,000 yr B.P.; 3

paroxysmal volcanic eruptions and formation of theSunda caldera and the east Lembang fault ca. 105,000

Ž .yr B.P.; 4 ongoing lacustrine sedimentationŽ .105,000–50,000 yr B.P.; 5 second phase of Plinian

and caldera-forming eruptions, and Bandung volcani-Ž .clastic fan development 50,000-35,000 yr B.P.; 6

high lake levels and lacustrine sedimentation, andformation of the west Lembang fault 35,000–20,000

Ž .yr B.P.; 7 small eruptions of Tangkuban Perahuvolcano and minor basin subsidence and sedimenta-tion, and minor fluvial erosion after 16,000 yr B.P.

The events mark the significance of the Sunda–Tangkuban Perahu volcanic centre during the LateQuaternary; the Sunda volcano collapsed into acaldera in which later the Tangkuban Perahu volcanodeveloped. Moreover, these eruptions controlled re-gional sedimentation and determined landform de-velopment in the great basin area. In the vicinity ofthe eruption centre, volcano-tectonic faulting formedthe conspicuous E–W Lembang fault that controlleddistribution of volcaniclastic sediments and the initi-ation of a new drainage system in the Lembang area.

4. Contribution to volcanology: interplay of con-struction and denudation processes throughouteruptive activity

Volcanic geomorphology can contribute to physi-Ž .cal volcanology through the a assessment of topo-

graphic effects on transport, erosion, and depositionŽof volcanogenic flows e.g., pyroclastic density cur-

. Ž .rents , b analysis of the relationships between erup-Ž .tion phenomenology and eruptive processes, and c

identification of the sources and climaticrtectonicconditions which govern the emplacement of vol-caniclastic deposits.

4.1. Complex interplay of eruptiÕe actiÕity with geo-morphic processes

4.1.1. Topographical effects on transport and em-placement of pyroclastic flows

Observations of recent or ongoing eruptions sug-gest that high energy relief may exert effects ontransport and deposition of primary pyroclastic de-posits, such as the flow-surge laid down by the blast

Žon 18 May 1980 at Mount St. Helens Fisher, 1990,. Ž1995 , and the block-and-ash flows at Unzen 1991–.1993 . The decoupling of pyroclastic currents is

accentuated by encounters with steep mountain ridgesŽ .of high relief Fisher, 1995 . In regions of rugged

topography, the height of barrier ridges, slopes an-gles and gradients of the ground surface greatlyinfluence the effectiveness of decoupling processesand flow directions. The topographical effect ofbreaks in slope on the flanks of stratocones con-tributes to the decoupling of two zones within a

Ž .pyroclastic flow e.g., at Mt. Merapi, 1994 : the


channelized, dense, gravity-driven part of the pyro-clastic flow and the unconfined, dilute and low-con-centration, turbulent part of the flow whose destruc-tive effects project far beyond the valley channels.The interactions of pyroclastic currents with topogra-phy include blocking, downslope drainage, formationof secondary pyroclastic flows in valleys, and devel-opment of dividing streamlines and decouplingŽ .Fisher, 1995 .

4.1.2. Erosion during waxing phases of pyroclasticflows-surges

Waxing phases of blasts and pyroclastic flows cancreate erosional landforms such as erosional furrowsformed during the 1980 eruptions of Mount St.

ŽHelens Kieffer and Sturtevant, 1988; Kieffer and.Simonds, 1995 . Field estimate for the erosion rate

by the 18 May 1980 lateral blast is 20.6 kg my2 sy1,corresponding to an erosion depth of nearly 1 m in atime of 30 s. The 7 August 1980 pyroclastic flowtriggered an erosion rate of 14.3 kg m2 sy1, corre-

Žsponding to 2 m of erosion in 5 min Kieffer and.Simonds, 1995 . Pyroclastic flows generated in the

19–20 April 1993 eruption of Lascar Volcano, Chile,Žproduced spectacular erosion features Sparks et al.,

.1997 . Exposed bedrock and boulders suffered se-vere abrasion, producing smoothed surface on coarsebreccias and striations and percussion marks onbedrock and large boulders. Erosional furrows devel-oped with wavelengths of 0.5–2 m and depth of0.1–0.3 m. Erosive features were produced whereflows accelerated through topographic restrictions orwhere they moved over steep slopes. Much of theerosive phenomena are attributed to lithic clastswhich segregated to the base of the flows. Theerosive features, distribution of lithic clasts and de-posit morphology indicate that the 1993 flows werehighly concentrated avalanches dominated by parti-cle interactions.

4.1.3. Monitored and measured processes of domegrowth and magma supply

Processes of dome growth can be observedthrough aerial photographs and satellite images, and,in some cases, measured through a real-time moni-

Žtoring network e.g., Mount St. Helens, Santiaguito:.Anderson et al., 1995; Unzen, Merapi, Monserrat . A

dacite dome at Unzen volcano, for example, grew intwo pulses, mainly exogenously when it was small

Ž 5 3 y1and the effusion rate was high 4=10 m day ;.May 1991 to Februrary 1993 , but endogenously

when the dome became large and the effusion rateŽ 4 3 y1declined 5=10 m day ; February 1993 to Au-.gust 1994 . The volume of magma erupted during

each pulse was 1.3=108 m3 and 0.6=108 m3,Ž .respectively Nakada et al., 1995 .

Magma supply rates are high on recently grownŽdomes Redoubt, 1989–1990: Miller, 1994; Unzen,

.1991–1994: Nakada et al., 1995 or present-dayŽgrowing domes Soufriere Hills, Monserrat, 1996–

.1998 . At Monserrat, a dome appeared in mid-November 1995 in the English’s Crater of Soufriere

ŽHills within 4 months of the eruption’s onset Young.et al., 1997 . The dome has grown endogenously and

exogenously like at Unzen, while dome collapses onthe NE flank have triggered many block-and-ashpyroclastic flows that were channeled in the Tarriver valley towards the East. Since October 1996, anew dome has filled in the explosion vent andpyroclastic flows have resumed, even towards theWest where the capital city, Plymouth, only 4 kmfrom the crater, is therefore evacuated. Since January1997, magma supply increased again to 2=105 m3

y1 5 3 y1 Žday up to as much as 7=10 m day Young.et al., 1997 . In February 1997, the volume of the

dome was 40=106 m3.

4.2. Interactions in construction and denudation pro-cesses

4.2.1. LaÕa flow and coastal processesLava flows entering seawater create a prograding

Žlava delta on the South coast of Hawaı island Mat-¨.tox and Mangan, 1997 . For nearly 14 years, pahoe-

hoe flows from Kilauea advance down the flank ofthe volcano from the Pu’u’O’o lava cone and form asystem of tubes that transport lava to the coastline.An average volume of 350,000 m3 dayy1 of lavawas fed through the tube system between 1986 and1994. During this time, 2 km2 of new land wasadded to the island. Two types of interactions were

Žobserved at the front of the delta Mattox and Man-.gan, 1997, Fig. 9 : open mixing of lava and seawater

when the complete collapse of lava bench seversactive lava tube; confined mixing conditions whenpartial collapse of lava bench submerges and frac-tures a portion of active lava tube.


Ž .Fig. 9. After Mattox and Mangan, 1997 their Fig. 4, modified . Hypothetical cross section of a lava delta and bench showing the location ofŽ .littoral hydrovolcanic explosions. A Hyaloclastites formed at the ocean entry build a submarine debris slope that is subsequently capped by

Ž .pahoehoe flows. Lava tubes on the bench can reside below or at sea level, due to continuous subsidence of the delta. B Profile of the frontŽ .of the delta immediately following a complete bench collapse. C Profile of the front of the delta immediately following a partial collapse

Žof the bench. Reprinted from Journal of Volcanology and Geothermal Research, 75, 1–17, Copyright 1997, with permission from Elsevier.Science .

4.2.2. EruptiÕe actiÕity, sea-leÕel changes, and ero-sion at seamounts, guyots, and shoaling Õolcanoes

Effects of sea-level change induced erosion havebeen observed at Palinuro, Italy, and effects of sub-marine eruptions were discovered in 1996 at Loihiseamount, the next-to-be born Hawaıan island, SE of¨

Ž .Kilauea Carlowicz, 1996 . Lithofacies associationsat shoaling or active subaerial island volcanoes canbe related to stages, early and late, of emergenceŽ .McPhie et al., 1993; McPhie, 1995 . Facies associa-tions are sensitive to proximity to source vents andto water depth of eruption and emplacement. Litho-facies associations bear on the variation in eruptionand fragmentation processes with respect to environ-ment, especially water depth. Dry eruptions are lim-

ited to subaerial and very shallow water settings, andmay be explosive or effusive. Autoclastic fragmenta-tion operates universally although abundant hyalo-clastite is restricted to subaqueous environments.Pillows are a hallmark of subaqueous lava eruptionor flow of subaerial lava into subaqueous settings.

4.2.3. Interaction of collapse and exhumation pro-cesses at caldera walls

Present-day caldera walls often present a complexassemblage of cliff surfaces of different ages, whichcan result from repeated collapses that exhumedearlier caldera cliffs and unconformities on the samecaldera rims. Geomorphological mapping at San-

Ž .torini Druitt and Francaviglia, 1992 shows that the


caldera wall in the north, NE, and east preservesevidence for three generations of cliff surface, those

Ž .of Minoan age ca 3.6 ka s17th century BC andtwo earlier generations of caldera: the Skaros calderaabout 76–54 ka and the Cape Riva caldera ca. 21 ka,which exhumes cliffs of the Skaros caldera. There-fore, field relationships are critical in unravelling thegeomorphology of Bronze-Age Santorini immedi-ately before the Minoan eruption, with a large diame-ter, a central island, and a probably flooded caldera,because the 21 ka caldera walls extend to present-daysea level at several locations.

4.2.4. Volcano–glacier interactionsIce-clad or snow-covered active volcanoes are

home to eruptions during which combined eruptive,glacial, and geomorphic processes lead to generationof primary and secondary sediment—water flows

Žtermed lahars or volcanic debris flows Major and.Newhall, 1989; Pierson et al., 1990; Thouret, 1990 .

Five historic eruptions at four snow-clad volcanoesŽTokachi–Dake, Nevado del Ruiz, Cotopaxi, and

.Mount St. Helens have demonstrated thatŽ .snowmelt-generated volcanic debris flows can: 1

5 3 y1 Ž .have peak discharges as large as 10 m s , 2y1 Ž .attain velocities as high as 20–40 m s , 3 mobi-

8 3 Ž .lize as much as 10 m of debris, and 4 travel moreŽ .than 100 km as debris flows in valleys draining the

Ž .volcanoes Pierson, 1995 . The risk to human lifefrom such large debris flows was tragically demon-strated in 1985 at Nevado del Ruiz volcano inColombia, where snowmelt-triggered lahars in threeof the volcano’s major drainage systems killed more

Žthan 23,000 people Pierson et al., 1990; Thouret,.1990 .

Attention has been drawn upon new types ofvolcaniclastic sediments and flows, such as the‘volcanic mixed avalanches’ from the November 13,

Ž1985 eruption of Nevado del Ruiz volcano Vande-.meulebrouck et al., 1993; Pierson and Janda, 1994 ,

Žand unusual ‘ice diamicts’ comprising clasts ofglacier ice and subordinate rock debris in a matrix of

.ice, snow, coarse ash, and frozen pore water em-placed during the December 15, 1989 eruption of

Ž .Redoubt volcano, AK Waitt et al., 1994 , and theŽ .1992 eruption of Mount Spurr, AK Waitt, 1995 .

Transient, mixed avalanches transformed to initial

‘snow slurry’ lahars, then to large dilute and smallconcentrated lahars in the Whangaehu catchment of

Ž .Ruapehu in 1995 Cronin et al., 1996, 1997 . TheseŽhybrid wet flows between conventional pyroclastic

.flows and conventional lahars are hazardous insofaras their reduced internal friction projects destructiveflows down valleys beyond the reach of dry pyro-clastic currents. Such deposits at snowclad volcanoesare geomorphically distinct, but they soon becomeextensively reworked and hard to recognize in thegeologic record.

Yet the processes involved in volcano–glacierinteractions remain not well understood. Perturba-tions on ice-clad active volcanoes record a variety ofprocesses, including rapid melting, snow and iceavalanching, surficial abrasion, and mechanical

Ž .scouring or gullying Thouret et al., 1995 . The lossof large volumes of snow and ice during eruptions

Ž .results mainly from 1 the passage of pyroclasticŽ .flows and surges or hot blasts on the glacier, 2 the

contact of subaerial lava flows or tephra with ice orŽ .snow, and 3 the eruptive or geothermal activity

which melts the bases of ice caps. Large volumes ofŽmeltwater released in a short time span e.g., 38.5–

44=106 m3 in 20–90 min at Nevado del Ruiz on.13 November 1985 imply a high melting rate and a

vigorous heat transfer from hot eruptive products tosnow and ice. Preliminary melting scenarios basedon vigorous deposition of hot debris on snow pointto a melting rate as high as 2 cm miny1. Mechanicalentrainment and comminution of snow and ice areimportant processes in releasing large volumes ofmeltwater that contribute to trigger lahars.

5. Contribution to sedimentology: significance ofvolcaniclastic sediments, flows, and facies models

Ž .Volcanic geomorphology can a identify sedi-mentary facies associations and facies models for

Ž .dynamic volcano-sedimentary systems, b establishcriteria for recognizing volcaniclastic deposits in oldvolcanic successions, and infer the role of climatic–

Ž .tectonic effects on transport and deposition, and canalyse the characteristics of sediment gravity flowsto determine relevant parameters for modelling theirbehaviour.


5.1. Volcaniclastic sediments and flows

The rapid growth and denudation of volcaniclandforms combine to cause serious erosion indrainage systems and supply huge volumes of sedi-ments through the catchments toward neighbouringlowlands. Because tectonism and volcanism areclosely associated at plate margins, and in someinstances within plates, knowledge about volcaniclas-tic sediments and rocks may be critical for tectonics,depositional and erosional processes, as well as forunderstanding the economic significance of mineral-ized and alteration zones in both ancient and modernsuccessions.

The importance of volcaniclastic sediments raisesŽ .at least four questions: 1 How do we identify

volcanogenic sediments and lithofacies associations?Ž .2 What are the types, transport processes, and

Ž .behaviour of volcanogenic flows? 3 How much andhow rapidly do rates of sedimentationrerosion fluc-

Ž .tuate? 4 How quickly do disturbed catchmentsrecover following large eruptions?

5.1.1. Distribution and significance of Õolcaniclasticsediments and facies models

Sedimentary processes exert a great influence onmodern volcanoes, while reworked volcanic rocksare volumetrically important and must be significantin the geological record.

Generalised facies models are based on the identi-fication and distribution of proximal, medial, anddistal, nonmarine volcaniclastic facies and sedimen-tary cycles triggered by large eruptions around stra-

Žtovolcanoes, such as Fuego, Guatemala Vessell and.Davies, 1981 in: Cas and Wright, 1987 . Facies are

features of a sedimentary unit portraying the pro-cesses of origin and source, and environment ofdeposition. Among all modern volcanoes, stratovol-canoes are very prone to mass-wastage because theyare high topographic features and they host greatvolumes of easily removed fragmental material. Theirgrowth is therefore reflected almost instantly in thesedimentary record of surrounding regions.

Composite volcanoes are complex dynamic vol-cano-sedimentary systems. Hackett and HoughtonŽ .1989 proposed a facies model for the composite

Ž .Ruapehu volcano New Zealand of Quaternary age,with moderate discharge rates of magma and a wettemperate climate, and whose products are subject to

rapid erosion and reworking. Long intervals are char-Žacterised by small but frequent eruptions 1943,

.1995–1996 , providing a continual supply of debristo the surrounding ring plain. Larger explosive erup-tions trigger major pulses of ring plain sedimenta-tion. The volcano can be divided into two parts: acomposite cone of 110 km3 in volume, surroundedby an equally voluminous ring plain. Cone-formingsequences are dominated by sheet- and autobrec-ciated-lava flows, which seldom reach the ring plain.The ring plain is built predominantly from the prod-ucts of explosive volcanism, both the distal primarypyroclastic deposits and the reworked material erodedfrom the cone. Much of the material entering the ringplain is transported by lahars, either generated di-rectly by eruptions or triggered by the high intensityrain storms which characterize the region. Ring plaindebris are reworked rapidly by concentrated andhyperconcentrated streams in pulses of rapid aggra-dation immediately following eruptions and moregradually in the longer intervals between eruptions.

5.1.2. Significance of continental Õolcanogenic sedi-mentation

The complex variations in the distribution of vol-caniclastic sediments have several sources. For ex-ample, the complex distribution of the volcaniclasticdeposits in an active rift can be caused by faulting,sub-basin development, sources of primary materials,

Žand local depositional environments Mathisen and.McPherson, 1991 . However, in his study on conti-

nental deposition of an extensional basin in SE Ari-Ž . Ž .zona, Smith 1994 Fig. 10 offers an evaluation of

continental sedimentation in response to climate,rather than tectonics. Covariance of climatic condi-

Žtions recorded in the pedogenic-carbonate isotope. Ždata and sedimentologic parameters sedimentation

.rate, channel geometry, and facies abundancestrongly suggests that climate can produce temporalvariations in sedimentological processes that haveheretofore been attributed only to tectonics. Hence,caution should apply in asserting tectonics as theonly significant long-time period influence on depo-sitional style in nonmarine basins.

Tephro-chronology and associated dating methodscan provide a framework for volcanism, tectonism,and paleoenvironmental reconstruction of basins sur-rounding a volcanic province. Tephro-chronology


Ž .Fig. 10. After Smith, 1994 his Fig. 14, caption modified . Summary diagram of paleoclimate interpretations and temporal variations inŽ .sedimentological characteristics within the chronologic control of the magnetic-polarity time scale. A Paleoclimatic interpretation of

Ž . Ž .paleosol-carbonate isotope data. B Sediment accumulation rates not adjusted for compaction averaged over time intervals defined by themagnetic polarity time scale. Bar widths reflect variations in rates determined from different stratigraphic sections and include uncertainties

Ž .in the exact position of reversals in each section, which depends on spacing of paleomagnetic-sample sites. C Channel geometries andabundance of coarse channel facies in eastern piedmont sections. The average and range of the percentages of sections composed ofsediment coarser than medium sand are shown over intervals defined by the polarity time scale to facilitate comparison to accumulation

Ž .rates. Ranges for all sections and averages are shown; only channel-tract sections SW, RR, DW, CR2 are represented for depositionalŽ .interval II. D Temporal range of hydromorphic-paleosol and pond-carbonate rocks related to shallow water tables. Reproduced with the

Ž .kind permission of the Geological Society of America GSA Copyright permission no. 21569 .

was used as a tool in the 2-Ma-long geomorphicŽhistory of ignimbrite plateaus in New Zealand Ken-

nedy, 1994; Alloway et al., 1995; Shane et al.,.1996 . Fifty-four tephra beds span the interval 2–0.6


Ma and provide an event frequency of 1r19 kaŽmuch higher than the ignimbrite frequency 1r100.ka . The tephra beds provide a framework for a

paleoenvironmental reconstruction of the Taupo Vol-Ž .canic Zone southern North Island . Volcaniclastic

transport routes from the TVZ to basins in the southand southeast, and through the site of present moun-tain ranges, supplied material to a terrestrial lowlandfore-arc area in the interval 1.64–0.7 Ma. Uplift anddeformation since 0.7 Ma have disrupted pale-odrainage routes, diverting them to the north and

Ž .southwest Shane et al., 1996 .

5.2. Type, characteristics, and origins of Õol-canogenic flows

Volcanogenic flows can be divided into pyroclas-Ž . Ž .tic primary , volcaniclastic secondary , and epiclas-

tic flows that erode, transport, and redeposit frag-Žmental sediments on and around volcanoes e.g., Cas

.and Wright, 1987 . Among volcanogenic flows, thedebris avalanches and lahars or volcanic debris flowshave drawn the attention of volcanologists, geomor-phologists, hydrologists, and sedimentologists sincethe 1980 eruption of Mount St. Helens. Many laharflows have been watched, filmed, and monitored,

Žespecially on active stratovolcanoes Mount St. He-.lens: Janda et al., 1981; Sakurajima, Unzen, etc. ,

and considerable progress has been achieved in the90s in understanding the processes, behaviour andrheology of lahar flows. After pyroclastic flows,lahars are the most deadly volcanic phenomena onactive and dormant volcanoes, projecting effects farbeyond the areas affected by the pyroclastic flows.

5.2.1. A spectrum of sediment graÕity flowsA simple but synthetic classification of mass

movements and flows on natural steep slopes hasŽ .been proposed by Coussot and Meunier 1996 as a

function of solid fraction and material type. Debrisflows occupy a field between landslides and hyper-concentrated flows and encompass mudflows andgranular flows on the base of cohesion and one ortwo-phase flow. Sediment gravity flows differ fromflood flows, based on greater velocities, greater im-pact forces, depositional record, and longer-term ef-fects. Debris flows transform into hyperconcentratedflows and streamflows and, with debris avalanches,are part of a spectrum of subaerial sediment gravity

flows, whose main characteristics are sediment con-Žcentration, deformation rate and velocity Pierson

and Costa, 1987; Scott, 1988; Smith and Lowe,.1991 .

5.2.2. Origins and characteristics of cohesiÕe andnoncohesiÕe debris flows

Ž .Following the pioneer study by Neall 1976 ,recent research incorporates post-1980 trends inrecognition of origins related to landslides, surges ofmeltwater from hot volcanic products on snow andice, failures of natural dams formed by volcanicflows, especially debris avalanches, and glacial out-

Ž .burst flows Scott and Sheridan, 1997 . The di-Ž .chotomy of cohesive muddy and noncohesive

Ž .granular lahars is illustrated through grain-size plotsŽ .Scott, 1988; Scott et al., 1995 . The textures andorigins of cohesive and noncohesive lahars are anal-ysed in terms of transformation of noncohesive la-hars to hyperconcentrated flows and streamflowsŽ .Pierson and Scott, 1985 , textural changes withintransformations, transition facies, and cause of dearth

Ž .of fine sediment clay and silt . The distinction hasbeen made between the probable syneruptivity of

Žnon-cohesive, granular flows e.g., from pyroclastic.flows vs. the possible non-eruptive origins of cohe-

sive lahars, e.g., from debris avalanche, such as the3 Ž3.8 km Osceola mudflow from Mount Rainier Scott

.et al., 1995; Vallance and Scott, 1997 .

5.3. Transport processes and behaÕiour of sedimentgraÕity flows

Interpretations of transport processes and be-Ž .haviour of debris flows Scott and Sheridan, 1997

Ž .are based on 1 observations on modern flows andŽ .on deposits of both ancient and modern flows, 2

grain-size analysis, texture and sedimentologicalŽcharacteristics critical diameter, phases and facies of

debris-flow deposits, clast support vs. matrix sup-. Ž . Ž .port , 3 sedimentary structures such as a inverse

Ž .and normal grading and b boundary structures, i.e.,sole layer, sheared boundary, dewatering structures,lamination and stratification, sharp contacts, and in-

Žclusion of fragile megaclasts Scott, 1988; Scott et. Ž .al., 1995 , 4 sedimentary fabric of debris-flow de-

posits, clast roundness and composition, lahar bulk-ing factors, and progressive downstream improve-ment in sorting, increase in sand and gravel, and


decrease in clay. These downstream progressions arecaused by incorporation of better sorted gravel and

Žsand e.g., Osceola mudflow at Mount Rainier: Val-.lance and Scott, 1997 .

Present-day debate bears on behaviour and em-placement of debris flow, i.e., en masse emplace-

Žment as opposed to incremental deposition like pro-gressive aggradation for emplacement of pyroclastic

.flows . Normal grading observed in the Osceolamudflow-deposits from Mount Rainier is best ex-plained by incremental aggradation of a flow wave,

Žcoarser grained at its front than at its tail Vallance.and Scott, 1997, Fig. 11 .

The results and questions to be solved are usedfor the paleohydrological study of sediment gravityflows and to outline potentially hazardous areas to beaffected by such flows, in particular by lahars: esti-mation of the cross-sectional area of flow, of thevelocity, discharge, extent and volume of flows, and

Žpreparation of maps of inundation areas e.g., Mount.Rainier: Scott et al., 1995; Scott and Sheridan, 1997 .

The distinction among cohesive lahar, noncohesivelahar, and debris avalanche is important for thepurpose of hazard assessment, because cohesive la-hars spread much more widely than noncohesivelahars that travel similar distances, and travel fartherand spread more widely than debris avalanches of

Ž .similar volume Scott and Sheridan, 1997 .

6. Process-oriented geomorphology: denudationrates and geomorphic impact of substantial erup-tions on volcanic landscapes

Volcanic geomorphology should be moreŽ .process-oriented through the a development of ac-

curate methods for measuring rates of geomorphicprocesses that shape ephemeral volcanic constructs,Ž .b identification of sources of material which con-tribute to sediment-delivery systems and of the fac-tors that control the sedimentary budget on slopesand in valley channels draining active volcanoes, andŽ .c measurement and comparison of the geomorphic

Ž . Ž .Fig. 11. From Vallance and Scott 1997, their fig. 13, caption modified . Schematic diagram of stage height versus time hydrograph formass flow at point in bottom of valley, and depositional sequence of flow illustrating incremental-deposition model for explaining normalgrading. Deposition illustrated for two stations 1 and 2 at heights h and h above valley bottom and with ultimate thicknesses t and t .1 2 1 2

Short dashed lines through hydrograph indicate uniform incremental deposition. Sense of motion of mass flow is to left. Reproduced withŽ .the kind permission of the Geological Society of America GSA Copyright permission no. 21569 .


impact on selected catchments and their hydrologicresponse before, during, and after eruptions. Hence,volcanic geomorphology should help to refine sedi-mentary and geomorphic parameters for the expo-nential decay model.

6.1. Denudation rates on Õolcanic landforms and onlandscapes

A central question in volcanic geomorphology ishow short- and long-term denudation rates of volca-noes have varied over time, in response to erosionalepisodes promoted by climate and sea-level changes,and tectonics.

6.1.1. On elementary Õolcanic landformsElementary volcanic landforms, like monogenetic

cinder cones and single lava flows, have a clearstarting time of geomorphic development, dependingon climatic and physiographic factors. The evolutionof tephra cones has long been the focus of morpho-

Žmetric studies e.g., Kieffer, 1971 in: Cas and Wright,.1987 . From statistical analysis on 38 cinder cones

with a maximum age of 3 Ma from the San Fran-Ž .cisco volcanic field in Arizona, Wood 1980 distin-

guished three different stages. With time and increas-ing diameter, cones show decreases in cone height,cone heightrcone basal diameter ratio and slope, butthe ratio of crater diameter-cone basal diameter doesnot appear to change. On the other hand, incisionrates in lava flow bedrocks measured in a variety ofvolcano-tectonic environments average 12.7 cm kay1

Ž .Righter, 1997 . They range from as low as 0.5–8cm kay1 in Hawaı to as high as 23–25 cm kay1 in¨

Ž .the Atenguillo Valley, Jalisco Mexico and even upto 30 cm kay1 in Utah.

6.1.2. On composite landformsŽ .Francis 1993 has summarized five stages in the

erosional history of a volcanic stratocone. Karatson´Ž .1996 made use of the unique features of the Neo-generQuaternary volcanic chain in the Carpathians:age progression is reflected well in degraded strato-volcanoes from south to north in a similar moderatecontinental climate. Based on a complex morphomet-ric analysis for 19 crater remnants dated from 11 Mato 0.4 Ma, 26 variables have been examined byregression and factor analyses: numerical values of

Ž .crater enlargment 109 mrMa , internal valleyŽ .growth 1.3 kmrMa , and average cone lowering

Ž .31.5 mrMa seem to answer the question of to whatextent the stratovolcanoes are degraded. The calcu-lated erosion rates seem reasonable in worldwidecomparison in moderate continental climate: the conelowering rate 31.5 m May1 fits well with globaldenudation rates inferred from other methods. Notsurprisingly, time is the most important factor thatexplains the morphometric characteristics by about

Ž40%. Adding two more factors, size-depth crater–.cone and the distance from erosion base level, the

stratovolcano morphology can be explained by 75%.

6.1.3. Long-term denudation rate in Õolcanic land-scapes

Short-term rates throw little light on the issue fordetermining how landscapes have evolved over enor-mous periods of time, which requires reliablechronological markers. In SE Australia, however, thewidespread preservation of Tertiary basalts through-out the highlands and adjacent lowlands offersconsiderable scope for measuring the processes ofwearing down and wearing back in the long-term

Ž .denudation of a highland mass Nott et al., 1996 .Both of these processes, dominance of scarp retreatvs. summit lowering in the denudation of a highlandmass, are found to be insignificant compared to therole of fluvial gorge extension over the last 30 m.y.Headward advancement of the Shoalhaven Gorgehas been occurring at approximately 15 times the

Žrate of major escarpment retreat 2500 m vs. 170.mrMa , 250 times the average rate of summit lower-

ing, and 500 times the rate of interfluve consump-tion. Over the long term, the highlands in this regionwill become considerably more dissected well beforethey decrease substantially in height or are narrowed.The conclusion also has important bearing uponmodels predicting isostasic rebound from assumedcharacter and rates of denudation.

6.2. Geomorphic impact of eruptions and sedimentdeliÕery on landscapes

Recent studies on volcanic landscapes affected byŽeruptions indicate that the geomorphic impact ero-

sion and sedimentation, and the subsequent recovery.of disturbed watersheds is complex, with an initial


stage of accelerated erosion followed by an exponen-tial decrease over a few years. Noteworthy, themeasured decreasing rate does not imply that erosionrates return to ‘normal’ pre-eruption rates.

6.2.1. Exponential decrease of the impact of eruptionon landscapes

Most of the observations and studies of erosionalprocesses are short-term and follow the first periodafter the eruption. Great volumes of clastic materialsrapidly flood sedimentary environments and are sup-plied directly to basins, with sediment delivery anderosion rates increasing by two to four orders ofmagnitude above the pre-eruption rates for non-af-

Ž .fected areas Swanson et al., 1983 . However, ero-sion and sedimentation processes in volcanic land-scapes affected by small to modest eruptions show arapid decline, e.g., from 25–100 mm ay1 in the first2 years to 1–5 mm ay1 within 5 years of theeruption of Mount St. Helens. Initial rates decreasewith increase of infiltration in the ash layer, develop-ment of a vegetation cover and a stable drainagesystem.

6.2.2. Long-term effect of small to moderate erup-tions

Following up the observations of SegerstromŽ .1950 on erosional and depositional processes thattook place on the Paricutin’s cone and lava field,

Ž .Inbar et al. 1994 emphasized the long-term effectand the recovery rates of the different landscapes 50

Ž .years after the eruption 1943–1953 . Three mainperiods of erosion were distinguished by Segerstrom:Ž .1 accelerated erosion between the beginning of theeruption in 1943 and the end of the rain period of

Ž .1944; 2 a deceleration period between 1944 andŽ .1952; and 3 a gradient deceleration in erosion rates

during the post eruption period until 1970, until theyreach the normal values for the area. However, theerosion rates as of 1990 are about 50% above nor-mal. The trend in the next period will be of slowerrates of erosion and may extend over decades orcenturies until a complete rehabilitation of the area.

In addition, rates of erosion and recovery willdepend strongly on the different landforms in thearea. On the cone and in the crater, erosion processesare slow as the vegetation cover grows rapidly. Mostof the areas within the 25 cm isopach at about 8 km

distance from the cone, are still covered with deepash and unvegetated, thus exposed to erosional pro-cesses for a long time. Floodplains are the mostactive area, where continuous floods during the rainyperiod add sediments or induce the incision of chan-nels along the edges of the lava fields. By contrast,lava fields are the least affected by erosional pro-cesses and most of the flows have a fresh appearanceafter 50 years. Thus, erosion processes and integra-tion of drainage systems from basaltic lava flows areslow and may take centuries or thousands of years,as no integrated drainage is found in historically

Žformed lava fields in different world climates Inbar.et al., 1994, 1995 .

6.2.3. Long-term sediment yield from impacted wa-tersheds are poorly known

The impact of non-varying watershed parameterson erosion processes could be assessed over theLate-glacial and Holocene periods in the small Lac

Ž 2Chambon watershed 39 km , Massif Central,.France , which exhibits forms inherited from the last

glacier extension in ancient mountain relief withoutŽ .tectonics Macaire et al., 1997 . Computation of

stored material volumes and sediment yield valuesfrom plutonic and volcanic source rocks over thepast 15,500 years in the lake Chambon watershedgave accurate information: the mean mechanical ero-

Ž .sion capacity by slope processes 16"6 m was 13Ž .times greater than by running water 1.2"0.3 m ,

but it developed over only a quarter of the watershedsurface. Fluctuations in sediment yield consisted of a

Ž2.5-fold increase during cold and dry climate Young.Dryas contrasting to a moderate decrease during

Ž .humid climates Pre-Boreal . A threefold increase inerosion over the last 1400 years shows the impact ofhuman-induced deforestation. Erosion rates as highas those computed over thousand of years can beattained over a few years only in volcanic water-

Ž 3.sheds affected by voluminous eruptions G10 km .

6.3. Dramatic response from disturbed catchments tosubstantial eruptions

Two cases are presented: the geomorphic effectsof the 1980 eruption of Mount St. Helens and the

Ž .first annual sediment budget 1980–1981 , and theextraordinary hydrologic response of the volcanic


landscape following the 1991 eruption of MountŽ .Pinatubo The Philippines .

6.3.1. Assessment of the 1980–1981 sediment budgetaround Mount St. Helens

Ž .Rosenfeld 1996 used a geomorphic approachinstead of a hydrologic approach to assess the sedi-ment budget around Mount St. Helens. First, poten-tial sediment sources were identified, based on aerialphotography and field investigations. Second, mea-surable landform units were categorized as well asnumerous measurable temporary storage sites andsediment sinks. Third, a geomorphic classification ofsurface materials and landform types was con-structed from post-eruption, repeated aerial photogra-phy. In addition, 65 control points were establishedand marked along the Toutle river valley, providingover 150 photogrammetric cross-sections of the con-stantly changing channel network.

Two types of sediment budget information wereŽ .obtained Rosenfeld, 1996 . The total sediment stor-

age was as high as 26=108 m3. The net sedimentyield was computed as 65=106 m3 for the first yearafter the eruption: two thirds of which were deliv-ered through erosion by the channel network at theexpense of the debris-avalanche deposit. The geo-morphic sources, volume, size, range and transportcharacteristics associated with storm events of differ-ent magnitudes were assessed. Besides lahars, themost active erosional processes were shallow masswasting, rill and bank erosion, and fillingrbreachingof detention ponds.

6.3.2. Protracted and worsening impact aroundPinatubo

Ž 3The response to large eruptions G10 km of.ejecta had not been documented prior to the study

of the Mount Pinatubo’s lahars. The short-term im-pact around Pinatubo has been severe, but the mid-term impact appears protracted and worsening.

6.3.2.1. Impact and lahar occurrence. The climacticexplosive eruption of Mount Pinatubo on June 15,1991, which erupted a total bulk volume of 8.4 to10.4 km3, deposited 5 to 6 km3 of abundant, loosepumiceous pyroclastic-flow deposits in the heads ofvalleys draining the volcano and about 0.2 km3 of

tephra on the volcano’s flanks that would later be theprimary source sediment for lahars. Numerous debrisflows and hyperconcentrated flows were triggeredduring and following 1991 and affected 8 major

Ž .drainages of Mount Pinatubo Pierson et al., 1997 .Ž .They were triggered by 1 monsoonal rainstorms,

sometimes enhanced by the passage of typhoonsŽ .farther to the north, 2 volcanically induced convec-

Ž .tive rainstorms over localized heat sources, and 3breakouts from debris dammed lakes. Lahars havebeen flowing into densely populated areas of centralLuzon over the past 8 years, taking a toll of lives,leaving more than 50,000 persons homeless, affect-ing more than 1,350,000 people in 39 towns and fourlarge cities, and causing enormous property lossesŽ 2 .G1000 km of prime agricultural land and social

Ž .disruption Janda et al., 1997 . Although the areasaffected by lahars have progressively expanded, thefrequency of lahar events has decreased and thenumber of impacted river systems had dwindled tofour in 1995, as source materials were graduallydepleted.

6.3.2.2. Sources of material and flow types. Laharsediment at Mount Pinatubo came from five distinct

Ž .sources: 1 coarse tephra dropped from the eruptionŽ .columns of the mid-June eruptions; 2 pyroclastic-

Ž .flow deposits; 3 fine-grained tephra from ash-clouddeposits, phreatic explosions, and eruptive events

Ž . Ž .postdating June 15; 4 1991 lahar deposits; and 5unconsolidated volcaniclastic deposits predating June

Ž .15 Pierson et al., 1997 . Lahar deposition occurredprimarily on low-gradient, coalescing alluvial fans15 to 50 km downstream from the caldera at the baseof the volcano, where deposit thicknesses generallyranged from 0.5 to 5 m. Total depositional volumeon the east-side alluvial fans in 1991 was about 0.38km3, which is almost one-third of the potential con-tributing volume from the source pyroclastic sedi-ments. Channelized lahars having peak discharges inthe order of 100 to 1000 m3 sy1 typically werenoncohesive pumiceous debris flows, some of whichtransformed to hyperconcentrated flows prior to finaldeposition. Flows range from turbulent, erosive hy-perconcentrated flows to viscous, usually laminar

Ždebris flows PHIVOLCS-DOST-IAVCEI, 1995;.Pierson et al., 1997 . Lahars in 1991 to 1994 wereŽ .predominantly hot 508C and steaming, fed by sedi-


ments eroded from the thick pyroclastic-flow de-posits filling valleys in the upper reaches of thewatersheds. Succeeding lahars in 1995–1996 werepredominantly cold, and consisted of increasinglyhigher proportion of older, pre-1991 eruption de-posits.

6.3.2.3. Sediment budget and yield. The sedimentbudget at Mount Pinatubo was evaluated by Pierson

Ž . Ž . 3et al. 1992 Fig. 12 . In 1992, 2.3 km of pyroclas-Ž .tic debris about 40% of the 1991 pyroclastic flows

were bound to be delivered to rivers as long-termlahar deposits. At the end of the 1994 rainy season,about 2.2 km3 has been already eroded from the1991 pyroclastic-flow deposits and deposited on thealluvial aprons at the foot of the volcanoŽ .PHIVOLCS-DOST-IAVCEI, 1995 . Owing to theextraordinary thickness of the accumulated pyroclas-tic debris, occurrence of destructive lahars is ex-pected to continue for several years. A first approxi-mation of the potential yearly sediment budget, basedon an exponential decay model was made by Pierson

Ž . Ž .et al. 1992 Fig. 13A and was one of the primaryinput towards constructing a lahar hazard map thatcan be utilized for long-term planning and rehabilita-tion of the Mount Pinatubo area. A decay rate inter-mediate between those of Mount St. Helens and

ŽMount Galunggung was chosen for Pinatubo Fig..13A,B . This exponential decay model is continu-

ously refined as additional information became avail-able through succeeding years, and lahar hazardmaps are adjusted accordingly.

Sediment yields set world records during the firstthree posteruption years: sediment yields in 1991were on the order of 1 million m3 kmy2 ay1, nearlyan order of magnitude greater than the maximumsediment yield computed following the May 18,

Ž1980, eruption of Mount St. Helens Pierson et al.,.1997 . In fact, the prodigious sediment yield from

Pinatubo’s upper and middle slopes and the sedimentstorage capacity in the adjoining lowlands are both

Ždiminishing, but at mismatched rates Janda et al.,.1997 . In general, sediment yields peaked early and

are decreasing rapidly in east-side watersheds, wherethe volume of 1991 pyroclastic-flow deposits is rela-tively low, deposits and streams are confined in afew steep-walled valleys, thin ash fall from sec-ondary explosions is common, and vegetation recov-

ery is fast. Sediment yields peaked later and aredecreasing slowly in west-side watersheds, wherepyroclastic-flow deposits are more voluminous, nu-merous small streams drain a broad, gently-sloping,unconfined pyroclastic apron, and vegetation recov-ery was initially low.

The prodigious sediment yield from Pinatubo hasalso several aftermaths in terms of watershed disrup-tion or piracy, channel avulsion, and blockage oftributaries. Blockage of tributaries at their confluencewith the main channel, either by lahars or secondarypyroclastic flows, formed temporary lakes and im-poundments. Floods and cold hyperconcentratedflows triggered by breaching of these naturally-dammed tributaries provided an additional hazard atMount Pinatubo. Unlike rain-induced lahars, theycan occur even in the absence of lahar-triggeringrainfall, and therefore limit the capability to warn

Žthreatened areas e.g., Pasig–Potrero in 1994:.PHIVOLCS-DOST-IAVCEI, 1995 .

6.3.2.4. Cartographic modelling of erosion in theSacobia’s catchment. Prolonged intense rainfall as-sociated with typhoons, and geomorphic accidents,such as secondary pyroclastic flow-induced streampiracy, and lake breakout events, can significantlyalter the expected annual sediment delivery rate, asobserved in the competing Sacobia–Abacan–Pasig–

ŽPotrero river systems PHIVOLCS-DOST-IAVCEI,.1995 . A cartographic modelling of erosion in pyro-

Žclastic-flow deposits of Mount Pinatubo Daag and.van Westen, 1996 encompasses the rapidly chang-

ing geomorphology of the Sacobia catchment on theeastern slope of the volcano before and during theeruption and for three consecutive years afterward.Emphasis was given to the importance of streamcapture as a result of erosion and secondary explo-sions. To quantify the volumes of pyroclastic-flowmaterial and annual erosion, five digital elevationmodels were prepared and analyzed using a GIS, andpre- and post-eruption geomorphological maps wereelaborated. A total volume of 1.78 km3 of pyroclas-tic flows deposited in 1991 in the Sacobia catchmentcovered an area of 24 km2. Erosion rates werecalculated to be in the range of 136–219 millionm3ra, that is about 5.6 to 9.1 million m3 kmy2 ay1.

The Pinatubo case study raises at least threeŽimportant observations Major et al., 1997; Newhall

()

J.-C.T

houretrE

arth-ScienceR

eÕiew

s47

199995

–131

124

Ž .Fig. 12. From Pierson et al., 1992 their fig. 6 . Schematic portrayal of transportation and distribution of pyroclastic material produced by the June 1991 eruptions of Mt.Pinatubo. Thickness of the flux arrows and deposits is proportional to the estimated material volume.


Ž . Ž .Fig. 13. A From Pierson et al., 1992 their fig. 5 . Exponential decay curves fitted to decreasing annual sediment accumulation rates forMount Galunggung and Mount St. Helens. A decay rate intermediate between those two is chosen for Mount Pinatubo. Initial first-year

Ž . Ž .sediment accumulation rate for Mount Pinatubo is estimated. B From PHIVOLCS-DOST-IAVCEI, 1995 their table 2 .

. Ž .and Punongbayon, 1997 : 1 heavy rainfall alonewas not responsible for generating the lahars in

Ž .1991; 2 geomorphic ‘accidents’ affecting water-shed and channels play a significant role in redistri-


bution, reerosion, and redeposition of sediments.These will foster more lahars for the next 5 to 10years, and more flooding beyond the alluvial fans

Ž .onto the densely populated plains; 3 beyond thegeomorphic impact of the devastating 1991 lahars,the subsequent lahars triggered by the seasonal mon-soon rains and other geomorphic ‘accidents’ had fargreater social and economic impact.

7. Conclusion: key problems and volcanic hazards

Several key problems in volcanic geomorphologyare still not adequately addressed. A sound classifica-tion of volcanic landforms must take into accountconstructional vs. erosional origin, single againstmultiple factors of growth, eruption styles, and typeand volume of magma and erupted material. Labora-tory experiments and modeling should be carried outon the tectonic effect of volcanic constructs on tec-tonic setting and magmatic system. Detailed geologicmapping, high-resolution geochronology, and com-positional data should be combined to determineeruptive and constructional episodes and the instan-taneous destruction forces that govern the life of acomplex volcano. Large-scale instability and genesis

Ž .of flank failures raise two questions: 1 what pro-cesses in addition to gravitation can trigger slope

Ž .instability on low-angle shield volcanoes? 2 whatfundamental structural causes might lead to an edi-fice collapse? Experiments corroborate observationsthat the edifice itself can contribute to the weaknessof its bedrock.

To contribute better to volcanology, a geomor-Ž .phologist should: 1 document the erosive effects of

pyroclastic flows and effects of high energy relief ontransport and deposition of primary pyroclastic de-

Ž .posits, 2 investigate the mechanical and thermalprocesses involved in snow and ice melting throughpyroclastic flows and surges that affect snowpack at

Ž .active volcanoes, and 3 determine the factors thatinfluence behaviour and emplacement of debris flowsŽen masse emplacement as opposed to incremental

.deposition .Volcanic geomorphology should answer four

Ž .questions regarding volcaniclastic sedimentation: 1how do we identify volcanogenic sediments and

Ž .lithofacies associations? 2 what are the transport

Ž .processes and behaviour of volcanogenic flows? 3how much and how rapidly do rates of sedimenta-

Ž .tionrerosion fluctuate? 4 how quickly do disturbedcatchments recover following large eruptions?

A process-oriented volcanic geomorphologyŽ . Ž .Thouret, 1992 should focus on a measurements ofrates of geomorphic processes that shape ephemeral

Ž .volcanic constructs, b factors that control the sedi-mentary budget on slopes and in channels draining

Ž .active volcanoes, and c parameters that refine theexponential decay model.

Our review emphasizes two major needs. First,Ždata from different sources ground-based observa-

tions, statistical analyses, laboratory experiments,.numerical models should be integrated, as increas-

ing numbers of studies attest to the close relationshipand complex interplay between eruptive and deposi-tional processes, character of deposits, and environ-mental factors. Second, rates of geomorphic pro-cesses acting at all scales on volcanoes and volcaniclandscapes need to be measured accordingly in orderto reinforce the process-oriented aspect of volcanicgeomorphology.

7.1. Increasing number of people at risk aroundhazardous Õolcanoes

At least 500 million people will be living underŽthe shadow of a volcano by the year 2000 Tilling

.and Lipman, 1993 . Twice this century, large townshave been laid waste in minutes by volcanic erup-

Žtions St Pierre, Martinique, in 1902 and Armero,.Colombia, in 1985 . Major population centers lie just

tens of kilometers from several large volcanoes withŽa likelihood of eruption during the next century e.g.,

Napoli near Vesuvius, Seattle–Tacoma near Mount.Rainier, Manila near Taal; Pyle, 1995 . The problem

is of wider significance, because ‘good science aloneŽwill not do the job of reducing volcano risk’ Tilling

.and Lipman, 1993 . Although there have been manyadvances in our understanding of factors leading tothe occurrence of eruptions during the past decades,the ability to predict volcanic events is still poor.Research needs range from fundamental investiga-tion of the causative processes to direct scientific andengineering studies to mitigate risk. Increasingly,policy decisions are being based on a formalized‘risk assessment’ that requires a scientifically based


probabilistic determination of the likelihood of anatural hazard event occurring.

Geomorphological surveys form a logical startingŽ .point for natural hazard zoning Verstappen, 1988 .

Process-oriented geomorphology foster primary in-put for quantitative reconstruction of recent volcanicactivity, and for the development of models used in

Ž .long-term planning Rosi, 1996 . Geomorphologycan contribute to risk assessment through two ap-

Ž .proaches Slaymaker, 1996 : geomorphic hazardŽzonation and composite risk zonation e.g., Mount

.Pinatubo: Nossin and Javelosa, 1996 . Geomorphichazards, both volcanic and non-volcanic, are identi-fied and analysed with the aid of satellite imageryand field survey. Geomorphic hazard domains areestablished according to the capacity of each hazardto affect geomorphic stability, the perceived peoplevulnerability, and the priority of geo-resource func-tion. A composite risk zonation, incorporatinggeomorphic mapping, geomorphic risk analysis andgeoresource priority, is calculated. Additional riskassessment and zonation requires the development of

Ž .a series of scenarios Blong, 1996 in which eruptionmagnitudes, hazard types, composite risk zonationindices, and the vulnerability of people and infras-tructure are adequately considered.

References

Alloway, B., Neall, V.E., Vucetich, C.G., 1995. Late QuaternaryŽ .post 28,000 yr B.P. tephrostratigraphy of northeast andcentral Taranaki, New Zealand. J. R. Soc. New Zealand 25Ž .4 , 385–458.

Anderson, S.W., Fink, J.H., Rose, W.I., 1995. Mount St. Helensand Santiaguito lava domes: the effect of short-term eruptionrate on surface texture and degssing processes. J. Volcanol.Geotherm. Res. 69, 105–116.

Beget, J.E., Kienle, J., 1992. Cyclic formation of debris avalanchesat Mount Augustine volcano. Nature 356, 701–704.

Blake, S., 1990. Viscoplastic models of lava domes. In: Fink, J.H.Ž .Ed. , Lava Flows and Domes, Emplacement Mechanisms andHazard Implications. IAVCEI Proceedings in Volcanology,Vol. 2. Springer, pp. 89–126.

Blong, R.J., 1996. Volcanic Hazards risk assessment. In: Scarpa,Ž .R., Tilling, R.I. Eds. , Monitoring and Mitigation of Volcanic

Hazards. Springer, pp. 675–698.Buonasorte, G., Ciccacci, S., De Rita, D., Fredi, P., Lupia Palmieri,

E., 1991. Some relations between morphological character-istics and geologic structure in the Vulsini Volcanic ComplexŽ .Northern Latium Italy . Z. Geomorph. N.F. Suppl.-Bd. 82,59–71.

Calvari, S., Coltelli, M., Neri, M., Pompilio, M., Scribano, V.,1994. The 1991–1993 Etna eruption: chronology and lavaflow-field evolution. Acta Vulcanol. 4, 1–14.

Carracedo, J.C., 1994. The Canary Islands: an example of struc-tural control on the growth of large oceanic-island volcanoes.

Ž .J. Volcanol. Geotherm. Res. 60 3–4 , 225–241.Carracedo, J.C., 1996a. A simple model for the genesis of large

gravitational landslide hazards in the Canary Islands. In:Ž .McGuire, W.J., Jones, A.P., Neuberg, J. Eds. , Volcano

Instability on the Earth and Other Planets. Geol. Soc. London,Spec. Publ. 110, 125–135.

Carracedo, J.C., 1996b. Morphological and structural evolution ofthe western Canary Islands: hotspot-induced three-armed riftsor regional tectonic trends? discussion–reply. J. Volcanol.Geotherm. Res. 72, 151–162.

Cas, R.A.F., Wright, J.V., 1987. Volcanic Successions, Ancientand Modern. Unwin Hyman, 528 p.

Cotton, C.A., 1944. Volcanoes as Landscape Forms. Whitcombeand Tombs Publ., Christchurch, 416 p.

Coussot, Ph., Meunier, M., 1996. Recognition, classification andmechanical description of debris flows. Earth-Sci. Rev. 40,209–227.

Crandell, D.R., 1988. Gigantic debris avalanche of Pleistoceneage from ancestral Mount Shasta volcano, California, anddebris-avalanche hazard zonation. U.S. Geol. Surv. Bull. 1861,32 pp.

Cronin, S.J., Neall, V.E., Lecointre, J.A., Palmer, A.S., 1996.Unusual ‘snow-slurry’ lahars from Ruapehu volcano, New

Ž .Zealand, September 1995. Geology 24 12 , 1107–1110.Cronin, S.J., Neall, V.E., Lecointre, J.A., Palmer, A.S., 1997.

Changes in Whangaehu river lahar characteristics during the1995 eruption sequence, Ruapehu volcano, New Zealand. J.Volcanol. Geotherm. Res. 76, 47–61.

Daag, A., van Westen, C.J., 1996. Cartographic modelling oferosion in pyroclastic-flow deposits of Mount Pinatubo,Philippines. ITC J. 2, 110–124.

Dam, M.A.C., Suparan, P., Nossin, J.J., Voskuil, P.G.A., 1996. Achronology for geomorphological developments in the greaterBandung area, West-Java, Indonesia. J. Southeast Asian Earth

Ž .Sci. 14 1–2 , 101–115.Ž .Decker, R.W., Wright, T.L., Stauffer, P.H. Eds. , 1987. Volcan-

ism in Hawaı. U.S. Geol. Surv. Prof. Pap. 1350, 2 Vols., 1665¨pp.

Druitt, T.H., Francaviglia, V., 1992. Caldera formation on San-torini and the physiography of the islands in the late BronzeAge. Bull. Volcanol. 54, 484–493.

ŽCarlowicz, M., 1996. Earthquake swarm heats up Loihi Fall. Ž .Meeting Preview . EOS Trans. Am. Geophys. Union 77 42 ,

405–406.Fisher, R.V., 1990. Transport and deposition of a pyroclastic

surge across an area of high relief. The 18 May 1980 eruptionof Mount St. Helens, Washington. Geol. Soc. Am. Bull. 102,1038–1054.

Fisher, R.V., 1995. Decoupling of pyroclastic currents: hazardsŽ .assessments. In: Ida, Y., Voight, B. Eds. , Models of Mag-

matic Processes and Volcanic Eruptions. J. Volcanol.Geotherm. Res., 66, 257–263.


Francis, P.W., 1993. Volcanoes. A planetary Perspective. OxfordUniv. Press, 443 pp.

Gabbianelli, G., Romagnoli, C., Rossi, P.L., Calanchi, N., 1993.ŽMarine geology of the Panarea–Stromboli area Aeolian

.Archipelago Southeastern Tyrrhenian sea . Acta Vulcanol. 3,11–20.

Glicken, H.X., 1991. Sedimentary architecture of large volcanic-Ž .debris avalanches. In: Fisher, R.V., Smith, G.A. Eds. , Sedi-

mentation in Volcanic Settings. SEPM, Vol. 45, pp. 99–106.Gregg, T.K.P., Fink, J.H., 1995. Quantification of submarine

lava-flow morphology through analog experiments. GeologyŽ .23 1 , 73–76.

Hackett, W.R., Houghton, B.F., 1989. A facies model for aQuaternary andesitic composite volcano: Ruapehu, NewZealand. Bull. Volcanol. 51, 51–68.

Harris, A.J.L., Blake, S., Rothery, D.A., 1997. A chronology ofthe 1991 to 1993 Mount Etna eruption using advanced veryhigh resolution radiometer data: implications for real-time

Ž .thermal volcano monitoring. J. Geophys. Res. 102 B 4 ,7985–8003.

Hasenaka, T., 1994. Size, distribution, and magma output rate forshield volcanoes of the Michoacan–Guanajuato volcanic fieldCentral Mexico. J. Volcanol. Geotherm. Res. 63, 13–31.

Hausback, B.P., Swanson, D.A., 1990. Record of prehistoricdebris avalanches on the North flank of Mount St. Helens

Ž .volcano, Washington. Geosci. Can. 17 3 , 142–145.Hildreth, W., Lanphere, M.A., 1994. Potassium–argon

geochronology of a basalt–andesite–dacite system: the MountAdams volcanic field, Cascade Range of southern Washing-ton. Geol. Soc. Am. Bull. 106, 1413–1429.

Inbar, M., Hubp, J.L., Ruiz, L.V., 1994. The geomorphologicalevolution of the Paricutin cone and lava flows Mexico 1943–

Ž .1990. Geomorphology 9 1 , 57–76.Inbar, M., Risso, C., Parica, C., 1995. The morphological devel-

opment of a young lava flow in the South Western Andes—Ž .Neuquen Argentina. Z. Geomorph. N.F. 39 4 , 479–487.

Iverson, R.M., 1995. Can magma-injection and groundwater forcescause massive landslides on Hawaıan volcanoes? In: Ida, Y.,¨

Ž .Voight, B. Eds. , Models of Magmatic Processes and Vol-canic Eruptions. J. Volcanol. Geotherm. Res., 66, 295–308.

Janda, R.J., Scott, K.M., Nolan, K.M., Mattison, H.A., 1981.Lahar movement, effects, and deposits. In: Lipman, P.W.,

Ž .Mullineaux, D.R. Eds. , The 1980 Eruptions of Mount St.Helens, Washington. U.S. Geol. Surv. Prof. Pap. 1250, 461–478.

Janda, R.J., Daag, A.S., Delos Reyes, P.J., Newhall, C.G., Pier-son, T.C., Punongbayan, R.S., Rodolfo, K.S., Solidum, R.U.,Umbal, J.V., 1997. Assessment and response to lahar hazardaround Mount Pinatubo, 1991 to 1993. In: Newhall, C.G.,

Ž .Punongbayan, R.S. Eds. , Fire and Mud: Eruptions and La-hars of Mt. Pinatubo, Philippines. University of WashingtonPress, pp. 107–139.

Kahle, A.B., Abrams, M.J., Abbott, E.A., Mouginis-Mark, P.J.,Realmuto, V.J., 1995. Remote sensing of Mauna Loa. MaunaLoa Revealed: Structure, Composition, History, and Hazards.Geophys. Monogr. 92, AGU, 145–169.

Karatson, D., 1996. Rates and factors of stratovolcano degradation´

in a continental climate: a complex morphometric analysis fornineteen NeogenerQuaternary crater remnants in the Carpa-thians. J. Volcanol. Geotherm. Res. 73, 65–78.

Kauahikau, J., Margriter, S., Lockwood, J., Trusdell, F., 1995.Application of GIS to the estimation of lava flow hazards onMauna Loa Volcano, Hawaı. In: Mauna Loa Revealed: Struc-¨ture, Composition, History, and Hazards. Geophys. Monogr.92, AGU, pp. 315–326.

Kennedy, N., 1994. New Zealand tephro-chronology as a tool ingeomorphic history of the c. 140 ka Mamaku ignimbriteplateau and in relating oxygen isotope stages. GeomorphologyŽ .9 2 , 97–115.

Kieffer, S.W., Sturtevant, B., 1988. Erosional furrows formedduring the lateral blast at Mount St. Helens, May 18, 1980. J.

Ž .Geophys. Res. 93 B 12 , 14793–14816.Kieffer, S.W., Simonds, C.H., 1995. Constraints on the rate of

erosion of high-speed volcanic flows. Volcanoes in TownŽ .Roma, September 1995 . Period. Mineral. 64, 209–211.

Lecuyer, F., Bellier, O., Gourgaud, A., Vincent, P.M., 1997.ŽTectonic control of the Tondano caldera North Celebes, In-

.donesia . C.R. Acad. Sci., Paris, ser. II 325, 607–613.´Lipman, P.W., 1995. Declining growth of Mauna Loa during the

last 100,000 years: rates of lava accumulation versus gravita-Ž .tional subsidence. In: Rhodes, J.M., Lockwood, J.P., Eds. ,

Mauna Loa revealed: Structure, Composition, History, andHazards. Geophys. Monogr. 92, AGU, pp. 45–78.

Lopez, D.L., Williams, S.N., 1993. Catastrophic volcanic col-lapse: relation to hydrothermal processes. Science 260, 1794–1796.

Macaire, J.J., Bossuet, G., Choqier, A., Cocirta, C., De Luca, P.,Dupis, A., Gay, I., Mathey, E., Guenet, P., 1997. Sedimentyield during Late Glacial and Holocene periods in the LacChambon watershed Massif Central, France. Earth Surf. Pro-cesses Landforms 22, 473–489.

Macdonald, G.A., 1972. Volcanoes. Prentice-Hall, NJ, 510 pp.Macdonald, K.C., Fox, P.J., Alexander, R.T., Pockalny, R., Gente,

P., 1996. Volcanic growth faults and the origin of Pacificabyssal hills. Nature 380, 125–129.

Ž .McGuire, W.J., Jones, A.P., Neuberg, J. Eds. , 1996. Volcanoinstability on the Earth and other planets. Geol. Soc. London,Spec. Publ. Vol. 110, 388 pp.

McKnight, S.B., Williams, S.N., 1997. Old cinder cone or youngcomposite volcano?: the nature of Cerro Negro Nicaragua.

Ž .Geology 25 4 , 339–342.McPhie, J., 1995. A Pliocene shoaling basaltic seamount: Ba

volcanic group at Rakiraki, Fiji. J. Volcanol. Geotherm. Res.64, 193–210.

McPhie, J., Doyle, M., Allen, R., 1993. Volcanic textures, aGuide to the Interpretation of Textures in Volcanic Rocks.CODES, University of Tasmania, 196 pp.

Major, J.J., Newhall, C.G., 1989. Snow and ice perturbationduring historical volcanic eruptions and the formation of la-hars and floods—a global review. Bull. Volcanol. 52, 1–27.

Major, J.J., Janda, R.J., Daag, A.S., 1997. Watershed disturbanceand lahars on the east side of Mount Pinatubo during themid-June 1991 eruptions. In: Newhall, C.G., Punongbayan,

Ž .R.S. Eds. , Fire and Mud: Eruptions and Lahars of Mt.


Pinatubo, Philippines. University of Washington Press, pp.895–918.

Marti, J., Ablay, G.J., Bryan, S., 1996. Comment on ‘The CanaryIslands: an example of structural control on growth of largeoceanic-island volcanoes’ by J. C. Carracedo. Discussion–comment. J. Volcanol. Geotherm. Res. 72, 143–149.

Masson, D.G., 1996. Catastrophic collapse of the volcanic islandof Hierro 15 ka ago and the history of landslides in the Canary

Ž .islands. Geology 24 3 , 231–234.Mathisen M.E., McPherson J.G., 1991. Volcaniclastic deposits:

implications for hydrocarbon exploration. In: Fisher, R.V.,Ž .Smith, G.A. Eds. , Sedimentation in volcanic settings. SEPM

Spec. Publ. 45, pp. 27–36.Mattox, T.N., Mangan, M.T., 1997. Littoral hydrovolcanic explo-

sions: a case study of lava–seawater interaction at KilaueaVolcano. J. Volcanol. Geotherm. Res. 75, 1–17.

Mendel, V., Sauter, D., 1997. Seamount volcanism at the superslow-spreading Southwest Indian Ridge between 578E and

Ž .708E. Geology 25 2 , 99–102.Merle, O., Borgia, A., 1996. Scaled experiments of volcanic

Ž .sprading. J. Geophys. Res. 101 B 6 , 13805–13817.Miller, T.P., 1994. Dome growth and destruction during the

1989–1990 eruption of Redoubt volcano. J. Volcanol.Geotherm. Res. 62, 197–212.

Moore, J.G., Clague, D.A., 1992. Volcano growth and evolutionof the island of Hawaı. Geol. Soc. Am. Bull. 104, 1471–1484.¨

Moore, J.G., Clague, D.A., Holcomb, R.T., Lipman, P.W., Nor-mak, W.R., Torresan, M.E., 1989. Prodigious submarine land-slides on the Hawaıan ridge. J. Geophys. Res. 94, 17465–¨17484.

Moore, J.G., Normak, W.R., Holcomb, R.T., 1994. Giant Hawaıanünderwater landslides. Science 264, 46–47.

Moore, J.G., Chadwick, W.W., 1995. Offshore geology of MaunaLoa and adjacent areas, Hawaı. In: Rhodes, J.M., Lockwood,¨

Ž .J.P., Eds. , Mauna Loa Revealed: Structure, Composition,History, and Hazards. Geophys. Monogr. 92, AGU, pp. 21–44.

Nakada, S., Miyake, Y., Sato, H., Oshima, O., Fujinawa, A.,1995. Endogenous growth of dacite dome at Unzen volcanoŽ . Ž .Japan , 1993–1994. Geology 23 2 , 157–160.

Neall, V.E., 1976. Lahars—Global Occurrence and AnnotatedBibliography, Publ. 5. Victoria Univ. Wellington, NewZealand, 18 pp.

Newhall, C.G., Punongbayon, R.S., 1997. Fire and Mud. TheEruption of Mt. Pinatubo, Philippines. University of Washing-ton Press, 1300 pp.

Nossin, J.J., Javelosa, 1996. Geomorphic risk zonation related toJune 1991 eruption of Mt Pinatubo, Luzon, Philippines. In:

Ž .Slaymaker, O. Ed. , Geomorphic Hazards. Wiley, pp. 69–94.Nossin, J.J., Voskuil, R.P.G., Dam, R.M.C., 1996. Geomorpho-

logic development of the Sunda volcanic complex, west JavaIndonesia. ITC Journal 2, 157–165.

Nott, J., Young, R., McDougall, I., 1996. Wearing down, wearingback, and gorge extension in the long-term denudation of ahighland mass: quantitative evidence from the Shoalhavencatchment, Southeast Australia. J. Geol. 104, 224–232.

Ollier, C.-D., 1988. Volcanoes. B. Blackwell, Oxford.Ollier, C., 1995. Tectonics and landscape evolution in southeast

Australia. Geomorphology 12, 37–44.

Owen, S., Segall, P., Freymuller, J., Miklius, A., Denlinger, R.,Arnadottir, T., Sako, M., Burgmann, R., 1995. Rapid deforma-tion of the south flank of Kilauea volcano Hawaı. Science 267,¨1328–1332.

Pain, C.F., Ollier, C.D., 1995. Inversion of relief: a component ofŽ .landscape evolution. Geomorphology 12 2 , 151–165.

Palacios, D., 1994. The origin of certain wide valleys in theŽ .Canary Islands. Geomorphology 9 1 , 1–18.

Peterson, D.W., Moore, R.B., 1987. Geologic history and evolu-tion of geologic concepts, Island of Hawaı. In: Decker, R.W.,¨

Ž .Wright, T.L., Stauffer, P.H. Eds. , Volcanism in Hawaı. U.S.¨Geol. Surv. Prof. Pap. 1350, Vol. 1, 149–189.

Ž .PHIVOLCS-DOST-IAVCEI CVS , 1995, International FieldWorkshop on Pinatubo Lahars and Watershed Processes, An-

Žgeles City, Phillipines, October 17–21, 1995, 35 pp. with.Abstracts .

Pierson, T.C., 1995. Flow characteristics of large eruption-tri-ggered debris flows at snow-clad volcanoes; constraints for

Ž .debris-flow models. In: Ida, Y., Voight, B. Eds. , Models ofMagmatic Processes and Volcanic Eruptions, J. Volcanol.Geotherm. Res. 66, 283–294.

Pierson, T.C., Costa, J.E., 1987. A rheologic classification ofsubaerial sediment–water flows. In: Costa, J.E., Wieczorek,

Ž .G.E. Eds. , Debris flowsrAvalanches: Process, Recognition,and Mitigation. Geol. Soc. Am., Rev. Eng. Geol. 7, 1–12.

Pierson, T.C., Janda, J.J., 1994. Volcanic mixed avalanches: adisaster eruption-triggered mass-flow process at snow-cladvolcanoes. Geol. Soc. Am. Bull. 106, 1351–1358.

Pierson, T.C., Scott, K.M., 1985. Downstream dilution of a lahar:transition from debris flow to hyperconcentrated stramflow.Water Resource Res. 21, 1511–1524.

Pierson, T.C., Janda, J.J., Thouret, J.C., Borrero, C.A., 1990.Perturbation and melting of snow and ice by the 13 November1985 eruption of Nevado del Ruiz, Colombia, and consequentmobilization, flow, and deposition of lahars. J. Volcanol.Geotherm. Res. 41, 17–66.

Pierson, T.C., Janda, R.J., Umbal, J.V., Daag, A.S., 1992. Imme-diate and long-term hazards from lahars and excess sedimenta-tion in rivers draining Mount Pinatubo, Philippines. U.S. Geol.Surv. Water-Resour. Invest. Rep. 92-4039, 37 pp.

Pierson, T.C., Daag, A.S., Delos Reyes, P.J., Regalado, M.T.,Solidum, R.U., Tubianosa, B.S., 1997. Flow and deposition ofposteruption hot lahars on the East side of Mount Pinatubo,July–October 1991. In: Newhall, C.G., Punongbayan, R.S.Ž .Eds. , Fire and Mud: Eruptions and Lahars of Mt. Pinatubo,Philippines. University of Washington Press, pp. 921–950.

Pyle, D.M., 1995. Volcanoes; reduction of urban hazards. NatureŽ . Ž .London 378 6553 , 134–135.

Ž .Rhodes, J.M., Lockwood, J.P., Eds. , 1995. Mauna Loa revealed:Structure, Composition, History, and Hazards. Geophys.Monogr. 92, AGU, p. 348.

Righter, K., 1997. High bedrock incision rates in the Atenguilloriver valley Jalisco, Western Mexico. Earth Surf. ProcessesLandforms 22, 337–343.

Robin, C., Eissen, J.-Ph., Monzier, M., 1993. Giant tuff cone andŽ12-km-wide associated caldera at Ambrym volcano Vanuatu

.New Hebrides Arc . J. Volcanol. Geotherm. Res. 55, 225–238.Rosenfeld, Ch.-L., 1996. Monitoring of geomorphic effects of the


1980 eruptions of Mount St. Helens, Washington, USA. Geo-Ž .journal 38 3 , 321–328.

Rosi, M., 1996. Quantitative reconstruction of recent volcanicactivity: a contribution to forecasting of future eruptions. In:

Ž .Scarpa, R., Tilling, R.I. Eds. , Monitoring and Mitigation ofVolcanic Hazards. Springer, pp. 631–674.

Rossi, M.-J., 1996. Morphology and mechanism of eruption ofŽ .postglacial shield volcanoes in Iceland. Bull. Volcanol. 57 7 ,

530–540.Rossi, M.-J., Gudmundsson, A., 1996. The morphology and for-

mation of flow-lobe tumuli on Icelandic shield volcanoes. J.Volcanol. Geotherm. Res. 72, 291–308.

Rowland, S.K., 1996. Slopes, lava flow volumes, and vent distri-butions on Volcan Fernandina, Galapagos Islands. J. Geophys.

Ž .Res. 101 B12 , 27657–27672.Scarth, A., 1994. Volcanoes. An Introduction. Univ. Coll. London

Press, 273 pp.Scott, K.M., 1988. Origins, behavior, and sedimentology of lahars

and lahar-runout flows in the Toutle–Cowlitz River system.U.S. Geol. Surv. Prof. Pap. 1447-A, 76 pp.

Scott, K.M., Vallance, J.W., Pringle, P.T., 1995. Sedimentology,behavior, and hazards of debris flows at Mount Rainier,Washington. U.S. Geol. Surv. Prof. Pap. 1547, 56 pp.

Scott, K.M., Sheridan, M.F., 1997. Gravity driven flows of vol-canic origin: avalanches, debris flows, lahars, and pyroclasticflows from dome collapse. IAVCEI General Assembly, Coursen83, Puerto Vallarta, 2 booklets, 25–26 January 1997, unpub-lished.

Segerstrom, K., 1950. Erosion studies at Paricutin, state of Mi-choacan Mexico. U.S. Geol. Surv. Bull. 965-A, 164 pp.

Shane, P.A.R., Black, T.M., Alloway, B.V., Westgate, J.A., 1996.Early to middle Pleistocene tephrochronology of North Island,New Zealand: implications for volcanism, tectonism, and pale-

Ž .oenvironments. Geol. Soc. Am. Bull. 108 8 , 915–925.Siebert, L., 1996. Hazards of large volcanic debris avalanches and

Ž .associated eruptive phenomena. In: Scarpa, Tilling, R.I. Eds. ,Monitoring and Mitigation of Volcano Hazards. Springer, pp.541–658.

Siebert, L., Glicken, H.X., Ui, T., 1987. Volcanic hazards fromBezymianny- and Bandaı-type eruptions. Bull. Volcanol. 49,¨435–459.

Singer, B.S., Thompson, R.A., Dungan, M.A., Feeley, T.C., Nel-son, S.T., Pickens, J.C., Brown, L.L., Wulff, A.W., Davidson,J.P., Metzger, J., 1997. Volcanism and erosion during the past930 k.y. at the Tatara–San Pedro complex Chilean Andes.

Ž .Geol. Soc. Am. Bull. 109 2 , 127–142.Ž .Slaymaker, O. Ed. , 1996. Introduction: Geomorphic Hazards.

Wiley, pp. 1–7.Smith, G.A., 1994. Climatic influences on continental deposition

during late-stage filling of an extensional basin, southeasternArizona. Geol. Soc. Am. Bull. 106, 1212–1228.

Smith, D.K., Cann, J.R., 1992. The role of seamount volcanism inŽ .crustal construction at the Mid-Atlantic Ridge 24–308N . J.

Ž .Geophys. Res. 97 B 2 , 1645–1658.Smith, G.A., Lowe, D.R., 1991. Lahars: volcano-hydrologic events

and deposition in the debris flow-hyperconcentrated flow con-

Ž .tinuum. In: Fisher, R.V., Smith, G.A. Eds. , Sedimentation inVolcanic Settings. SEPM Spec. Publ., Vol. 45, pp. 59–70.

Smith, D.K., Humphris, S.E., Tivey, M.A., Cann J.R., 1997.Viewing the morphology of the Mid-Atlantic Ridge from anew perspective. EOS, Trans. Am. Geophys. Union 78, pp.26, 265, 269.

Smoot, N.C., 1995. Mass wasting and subaerial weathering inguyot formation: the Hawaıan and Canary Ridges as examples.¨Geomorphology 14, 29–41.

Smoot, N.C., King, R.E., 1993. Three-dimensional secondarysurface geomorphology of submarine landslides on Northwest

Ž .Pacific plate guyots. Geomorphology 6 2 , 151–173.Sohn, Y.K., 1996. Hydrovolcanic processes forming basaltic tuff

rings and cones on Cheju Island Korea. Geol. Soc. Am. Bull.Ž .108 10 , 1199–1211.

Sparks, R.S.J., Gardeweg, M.C., Calder, E.S., Matthews, S.J.,1997. Erosion by pyroclastic flows on Lascar volcano Chile.Bull. Volcanol. 58, 557–565.

Stevens, N.F., Murray, J.B., Wadge, G., 1997. The volume andshape of the 1991–93 lava-flow field at Mount Etna Sicily.

Ž .Bull. Volcanol. 58 6 , 449–454.Stoopes, G.R., Sheridan, M.F., 1992. Giant debris avalanches

from the Colima volcanic complex, Mexico: implication forŽ .long-runout landslides )100 km and hazard assessment.

Geology 20, 299–302.Swanson, F.J., Collins, B., Dunne, T., Wicherski, B.P., 1983.

Erosion of tephra from hillslopes near Mount St. Helens andother volcanoes. In: Symp. on Erosion Control in VolcanicAreas, Seattle, July 1982. Ibaraki, Japan, Public Works Re-search Institute, pp. 183–221.

Tanguy, J.-C., Kieffer, G., Patane, G., 1996. Dynamics, lavavolume and effusion rate during the 1991–1993 eruption ofMount Etna. J. Volcanol. Geotherm. Res. 71, 259–265.

Thouret, J.C., 1990. Effects of the 13 November 1985 eruption onthe ice cap and snow pack of Nevado del Ruiz, Colombia. J.Volcanol. Geotherm. Res. 41, 177–201.

Thouret, -J.-C. 1992. Des paleovolcans aux strato-volcans actifs.´Ž .In: Lageat, Y., Thouret, J.-C. Eds. , Geomorphologie vol-´

canique. Bull. Assoc. Geol. Fra., Vol. 4, pp. 326–366.Thouret, J.-C., Vandemeulebrouck, J., Komorowski, J.C., Valla,

F., 1995. Volcano-glacier interactions: field survey, remoteŽsensing and modelling—a case study Nevado del Ruiz

.Colombia . Steepland Geomorphology, Vol. 5. Wiley, pp.63–88.

Tilling, R.I., Lipman, P.W., 1993. Lessons in reducing volcanorisk. Nature 364, 277–280.

Tilling, R.I., Dvorak, J.J., 1993. Anatomy of a basaltic volcano.Ž .Nature London 363, 125–133.Ž . ŽTrigila, R., Ed. , 1995. The Volcano of the Alban Hills A

contribution to: geomorphology, stratigraphy and volcano-tectonics, mineralogy, petrology, geochronology, crustal seis-mology, gravimetry, hydrogeology, fluid geochemistry and

.thermalism, geochemical monitoring, volcanic hazard , Roma,283 pp.

Vallance, J.W., Scott, K.M., 1997. The Osceola mudflow fromMount Rainier: sedimentology and hazard implications of a


Ž .huge clay-rich debris flow. Geol. Soc. Am. Bull. 109 2 ,143–163.

Vallance, J.W., Siebert, L., Rose, W.I. Jr., Giron, J.R., Banks,N.G., 1995. Edifice collapse and related hazards in Guatemala.

Ž .In: Ida, Y., Voight B. Eds. , Models of Magmatic Processesand Volcanic Eruptions. J. Volcanol. Geotherm Res. 66, 337–355.

Van Wyk de Vries, B., Francis, P.W., 1997. Catastrophic collapseat stratovolcanoes induced by gradual volcano spreading. Na-ture 387, 387–390.

Van Wyk de Vries, B., Merle, O., 1996. The effect of volcanicŽ .constructs on rift fault patterns. Geology 24 7 , 643–646.

Vandemeulebrouck, J., Thouret, J.-C., Dedieu, J.-P., 1993. Recon-naissance par teledetection des produits eruptifs et des lahars´ ´ ´ ´sur et autour de la calotte glaciaire du Nevado del Ruiz

Ž .Colombie. Bull. Soc. Geol. France 164 6 , 795–806.´Verstappen, H., 1988. Geomorphological surveys and natural

hazard zoning, with special reference to volcanic hazards incentral Java. Z. Geomorphol. N.F., Sppl.-Bd. 68, 81–101.

Voight, B., Janda, R.J., Glicken, H. ougl, Douglss, P.M., 1983.Nature and mechanics of the Mount St. Helens rockslide-avalanche of 18 May, 1980. Geotechnique 33, 243–273.´

Wadge, G., Francis, P.W., Ramirez, C.F., 1995. The SocompaŽ .collapse and avalanche event. In: Ida, Y. Voight, B. Eds. ,

Models of Magmatic Processes and Volcanic Eruptions. J.Volcanol. Geotherm. Res. 66, 309–336.

Waitt, R.B., 1995. Hybrid wet flows formed by hot pyroclastsinteracting with snow during the 1992 eruptions of Crater

Ž .Peak, Mount Spurr volcao, Alaska. In: Keith, T.E.C. Ed. ,The 1992 eruptions of Crater Peak vent, Mount Spurr Vol-cano, Alaska. U.S. Geol. Surv. Bull. 2139, 107–118.

Waitt, R.B., Gardner, C.A., Pierson, T.C., Major, J.J., Neal, C.A.,

1994. Unusual ice diamicts emplaced during the December 15,1989 eruption of Redoubt volcano Alaska. J. Volcanol.Geotherm. Res. 62, 409–428.

Wood, C.A., 1980. Morphometric analysis of cinder cone degra-dation. J. Volcanol. Geotherm. Res. 8, 137–160.

Young, S. et al., 1997. The ongoing eruption in Monserrat.Science 276, 371–372, Together with the Monserrat VolcanoObservatory Team.

Jean-Claude Thouret is Professor at theŽUniversite Blaise Pascal Clermont-Fer-´

.rand II and research geomorphologist atUMR 6524 ‘Magmas and Volcanoes’ ofthe French CNRS since 1990. He earnedthe highest University’s degree from theEcole Normale Superieure de Saint-´

Ž .Cloud Paris in 1975 and has been As-sistant Professor at the Universite Joseph´

Ž .Fourier Grenoble I, 1975–1990 . Heearned his PhD in 1988, that dealt withthe volcanic and glacial geomorphology

of the Colombian Cordillera Central, focusing on the reknownNevado del Ruiz volcanic massif. He has been involved ingeomorphological and volcanological studies in Colombia, USAŽ . Ž .Mt. Hood , Indonesia Mt. Merapi, Kelud and Galunggung , and

Ž .Peru Nevado Sabancaya, El Misti, Huaynaputina, and Ubinas .Ž .At present, he is on leave at IRD formerly ORSTOM , the French

Institute for Research and Development, and has been working onvolcanoes in southern Peru for the past four years, in the frame ofa scientific and cooperative venture with the Geophysical Institute

Ž .of Peru Lima .

Documents

Thouret 1999 Volcanic Geomorphology-An Overview