Structural analysis of the early stages of catastrophic stratovolcano flank-collapse using analogue models

Many major volcanic flank collapses involve the failure of low-angle strata in or under the edifice. Such failures produce voluminous, destructive debris avalanches that are a major volcanic hazard. At Socompa, Las Isletas-Mombacho and Parinacota volcanoes, field studies have shown that during catastrophic flank collapse a significant segment of their substrata was detached and expelled from beneath the volcanic edifice and formed a mobile basal layer on which the sliding flanks were transported. Previous studies have proposed that gravitational flank spreading was likely involved in the onset of sudden substrata failure. The early stages of this particular type of flank collapse can be modelled under laboratory conditions using analogue models. This allows us to study the development of structures accommodating early deformation of the sliding flank during catastrophic collapse. In the experiments, the detached substratum segment (low-viscosity basal layer) was modelled with a silicone layer, and the overlying stratovolcano with a layered sand cone. The first structure developed in the models is a graben rooted in the low-viscosity basal layer. This graben forms the limits of the future avalanche-amphitheatre and divides the sliding flank into a ‘toreva’ domain (upper sliding flank) and a ‘hummock’ domain (lower sliding flank). These domains display distinctive structural patterns and kinetic behaviour. Normal faults develop in the toreva domain and inside the graben, while the hummock domain is characterised by transtensional structures. The hummock domain also over-thrusts the lower amphitheatre sides, which allows subsequent sideways avalanche spreading. Measurements show that horizontal speeds of the hummock domain are always higher than that of the toreva domain during model collapse. The main role played by the low-viscosity basal layer during this type of collapse is to control the size, shape and structural complexity of the sliding flank; it also transmits mass and momentum from the toreva to the hummock domain.     

Hydrothermal calderas

The standard model of caldera formation is related to the emptying of a magma chamber and ensuing roof collapse during large eruptions or subsurface withdrawal. Although this model works well for numerous volcanoes, it is inappropriate for many basaltic volcanoes (with the notable exception of Hawaii), as these have eruptions that involve volumes of magma that are small compared to the collapse. Many arc volcanoes also have similar oversized depressions, such as Poas (Costa Rica) and Aoba (Vanuatu). In this article, we propose an alternative caldera model based on deep hydrothermal alteration of volcanic rocks in the central part of the edifice. Under certain conditions, the clay-rich altered and pressurized core may flow under its own weight, spread laterally, and trigger very large caldera-like collapse. Several specific mechanisms can generate the formation of such hydrothermal calderas. Among them, we identify two principal modes: mode 1: ripening with summit loading and flank spreading and mode II: unbuttressing with flank subsidence and flank sliding. Processes such as summit loading or flank subsidence may act simultaneously in hybrid mechanisms. Natural examples are shown to illustrate the different modes of formation. For ripening, we give Aoba (Vanuatu) as an example of probable summit loading, while Casita (Nicaragua) is the type example of flank spreading. For unbuttressing, Nuku Hiva Island (Marquesas) is our example for flank subsidence and Piton de la Fournaise (La Réunion) is our example of flank sliding. The whole process is slow and probably needs (a) at least a few tens of thousands of years to deeply alter the edifice and reach conditions suitable for ductile flow and (b) a few hundred years to achieve the caldera collapse. The size and the shape of the caldera strictly mimic that of the underlying weak core. Thus, the size of the caldera is not controlled by the dimensions of the underlying magma reservoir. A collapsing hydrothermal caldera could generate significant phreatic activity and trigger major eruptions from a coexisting magmatic complex. As the buildup to collapse is slow, such caldera-forming events could be detected long before their onset.     

Edifice and substrata deformation induced by intrusive complexes and gravitational loading in the Mull volcano (Scotland)

It is likely that the structure of a volcanic edifice can be significantly modified by deformation caused by large, shallow intrusions. Such deformation may interact with that caused by volcano loading. We explore such intrusion-related and loading-related deformation with field evidence and analogue models. To do this we have chosen the eroded Palaeogene Mull volcano (Scotland) that had a major edifice, has well exposed intrusions and significant deformation. There are thin Mesozoic sedimentary rocks forming ductile layers below the volcano, but their thickness is insufficient to allow the gravitational spreading of the volcanic edifice, especially when considering that a thick lava pile covers them. Thus intrusive push may have been the driving force for deformation. The Mull activity migrated toward the northwest, forming three successive intrusive complexes (Centres 1, 2 and 3). Our detailed fieldwork reveals that deformation due to these was accommodated on three levels; along thrust planes in lava sequences, along a décollement located in a thin clay-rich sediment succession and in basement schists. A relative chronology has been established between different groups of structures using dyke and sill cross-cutting relationships. Centre 1 is surrounded by a fold and thrust belt leading to radial expansion. In contrast, Centre 2 and 3 are connected to thrusts located to the south and east, bounded by strike-slip faults, leading to expansion to the southeast. The migration of centres and the directed sliding of the edifice may be related to the presence to the southeast of low-resistance Dalradian basement that failed significantly during growth of Centres 2 and 3. To study the observed relationships we have carried out scaled analogue models. Models are made with fine powder intruded by a viscous magma analogue. The models show an intimate relationship between intrusion growth, uplift of the volcano and subsequent flank sliding. The structures produced can be compared with Mull and suggest that the Centre 1 thrust belt probably formed following edifice gravitational sliding as a consequence of the uplift associated with Centre 1 formation. Centre 2 and 3 are responsible for the sector sliding of the edifice flank toward the southeast as the magmatic complex became more asymmetric. The features observed at Mull and in the models are similar to those seen on active volcanoes, such as Etna, providing a structural framework for their deformation and evolution.     

The influence of edifice slope and substrata on volcano spreading

Gravitational volcano spreading is caused by flow of weak substrata due to volcanic loading, and is now a process known to affect many edifices. The process produces extension in the upper edifice, evidenced by gräben and normal faults, and compression at the base, seen in strike–slip faults and thrusts. Where spreading is identified, host volcanoes have a range of fault densities, variable rift and gräben shapes, and different degrees of structural asymmetry. Previous studies have suggested a link between edifice shape and structure and the proportion of brittle to ductile material in the substrata or lower edifice. We study this link using refined sand cone analogue models standing on a brittle–ductile/sand–silicone substrata. Two scenarios have been investigated, the first mainly represents oceanic volcanoes with a ductile layer within the edifice (type I), where there is an outer ductile free surface. The second represents most continental volcanoes that have ductile substrata (type II). We apply the model results to natural examples and develop quantitative relationships between slope, brittle–ductile ratio fault density, spreading rate and structural style. Displacement fields calculated from stereophotogrammetry show significant differences between different slope models. We find that more faults are produced when the cone is initially steeper, or when the brittle substratum is thinner. However, the effect of the brittle layer dominates over that of slope. The strike–slip movements are found to be an essential feature in the spreading mechanism and the gräben are in fact transtensional features. Strike–slip and graben faults make a conjugate flower pattern. The structures produced are well-organised for type II edifices, but they are poorly organised for type I models. Type I models represent good analogues for oceanic volcanoes that are commonly affected by large slumps bounded by an extensional zone and lack of well-formed sector gräben. The well-observed connection between oceanic volcano rifts and large landslide-slumps is confirmed to be a consequence of spreading.     

The 1998 debris avalanche at Casita volcano, Nicaragua — investigation of structural deformation as the cause of slope instability using remote sensing

During Hurricane Mitch in 1998, a debris avalanche occurred at Casita volcano, Nicaragua, resulting in a lahar that killed approximately 2500 people. The failure that initiated the avalanche developed at a pre-existing cliff, part of the headwall of a gravitational slide of approximately 1.8 km² in plan view that cuts the southern flank of the volcano. Structural analysis, primarily based on a high-resolution DEM, has shown that this slide is caused by edifice deformation. Casita's eastern side is spreading radially outwards, forming a convex–concave profile and steepening original slopes. This deformation is possibly facilitated by millennia of persistent hydrothermal alteration of the volcano's core. The gravity slide has some typical features of smaller slumps, such as steep headwalls, an inner flatter area and a pronounced basal bulge fronted by thrusts. The headwall is the source of the 1998 avalanche, as well as several previous mass movements. Edifice deformation has led to extensive fracturing of the hydrothermally altered andesitic source rock, increasing instability further. Field evidence indicates that the gravity slide is still actively deforming, and with steep headscarps remaining, the hazard of future avalanches is increasing. The analysis presented here shows how small but highly damaging landslides can occur during the deformation of a volcanic edifice. We show that identification of instability is possible with remote sensing data and minimal reconnaissance work, implying the possibility of similar efficient and cost-effective analysis at other volcanoes known to host extensive hydrothermal systems. We demonstrate this with a simple structural analysis of two similar stratovolcanoes, Orosí (Costa Rica) and Maderas (Nicaragua).     

Rift zone reorganization through flank instability in ocean island volcanoes: an example from Tenerife, Canary Islands

The relationship between rift zones and flank instability in ocean island volcanoes is often inferred but rarely documented. Our field data, aerial image analysis, and ⁴⁰Ar/³⁹Ar chronology from Anaga basaltic shield volcano on Tenerife, Canary Islands, support a rift zone—flank instability relationship. A single rift zone dominated the early stage of the Anaga edifice (~6–4.5 Ma). Destabilization of the northern sector led to partial seaward collapse at about ~4.5 Ma, resulting in a giant landslide. The remnant highly fractured northern flank is part of the destabilized sector. A curved rift zone developed within and around this unstable sector between 4.5 and 3.5 Ma. Induced by the dilatation of the curved rift, a further rift-arm developed to the south, generating a three-armed rift system. This evolutionary sequence is supported by elastic dislocation models that illustrate how a curved rift zone accelerates flank instability on one side of a rift, and facilitates dike intrusions on the opposite side. Our study demonstrates a feedback relationship between flank instability and intrusive development, a scenario probably common in ocean island volcanoes. We therefore propose that ocean island rift zones represent geologically unsteady structures that migrate and reorganize in response to volcano flank instability.Editorial responsibility: T. DruittThis revised version was published online in February 2005 with typographical corrections and a changed wording.     

Analogue models of the effect of long-term basement fault movement on volcanic edifices

Long-term fault movement under volcanoes can control the edifice structure and can generate collapse events. To study faulting effects, we explore a wide range of fault geometries and motions, from normal, through vertical to reverse and dip-slip to strike-slip, using simple analogue models. We explore the effect of cumulative sub-volcanic fault motions and find that there is a strong influence on the structural evolution and potential instability of volcanoes. The variety of fault types and geometries are tested with realistically scaled displacements, demonstrating a general tendency to produce regions of instability parallel to fault strike, whatever the fault motion. Where there is oblique-slip faulting, the instability is always on the downthrown side and usually in the volcano flank sector facing the strike-slip sense of motion. Different positions of the fault beneath the volcano change the location, type and magnitude of the instability produced. For example, the further the fault is from the central axis, the larger the destabilised sector. Also, with greater fault offset from the central axis larger unstable volumes are generated. Such failures are normal to fault strike. Using simple geometric dimensionless numbers, such as the fault dip, degree of oblique motion (angle of obliquity), and the fault position, we graphically display the geometry of structures produced. The models are applied to volcanoes with known underlying faults, and we demonstrate the importance of these faults in determining volcanic structures and slope instability. Using the knowledge of fault patterns gained from these experiments, geological mapping on volcanoes can locate fault influence and unstable zones, and hence monitoring of unstable flanks could be carried out to determine the actual response to faulting in specific cases.     

Formation of caldera periphery faults: an experimental study

Changing stresses in multi-stage caldera volcanoes were simulated in scaled analogue experiments aiming to reconstruct the mechanism(s) associated with caldera formation and the corresponding zones of structural weakness. We evaluate characteristic structures resulting from doming (chamber inflation), evacuation collapse (chamber deflation) and cyclic resurgence (inflation and deflation), and we analyse the consequential fault patterns and their statistical relationship to morphology and geometry. Doming results in radial fractures and subordinate concentric reverse faults which propagate divergently from the chamber upwards with increasing dilation. The structural dome so produced is characterised bysteepening in the periphery, whereas the broadening apex subsides. Pure evacuation causes the chamber roof to collapse along adjacent bell-shaped reverse faults. The distribution of concentric faults is influenced by the initial edifice morphology; steep and irregular initial flanks result in a tilted or chaotic caldera floor. The third set of experiments focused on the structural interaction of cyclic inflation and subsequent moderate deflation. Following doming, caldera subsidence produces concentric faults that characteristically crosscut radial cracks of the dome. The flanks of the edifice relax, resulting in discontinuous circumferential faults that outline a structural network of radial and concentric faults; the latter form locally uplifted and tiltedwedges (half-grabens) that grade into horst-and-graben structures. This superimposed fault pattern also extends inside the caldera. We suggest that major pressure deviations in magma chamber(s) are reflected in the fault arrangement dissecting the volcanoflanks and may be used as a first-order indication of the processes and mechanisms involved in caldera formation.     

Relationships between volcano gravitational spreading and magma intrusion

Volcano spreading, with its characteristic sector grabens, is caused by outward flow of weak substrata due to gravitational loading. This process is now known to affect many present-day edifices. A volcano intrusive complex can form an important component of an edifice and may induce deformation while it develops. Such intrusions are clearly observed in ancient eroded volcanoes, like the Scottish Palaeocene centres, or in geophysical studies such as in La Réunion, or inferred from large calderas, such as in Hawaii, the Canaries or Galapagos volcanoes. Volcano gravitational spreading and intrusive complex emplacement may act simultaneously within an edifice. We explore the coupling and interactions between these two processes. We use scaled analogue models, where an intrusive complex made of Golden syrup is emplaced within a granular model volcano based on a substratum of a ductile silicone layer overlain by a brittle granular layer. We model specifically the large intrusive complex growth and do not model small-scale and short-lived events, such as dyke intrusion, that develop above the intrusive complex. The models show that the intrusive complex develops in continual competition between upward bulging and lateral gravity spreading. The brittle substratum strongly controls the deformation style, the intrusion shape and also controls the balance between intrusive complex spreading and ductile layer-related gravitational spreading. In the models, intrusive complex emplacement and spreading produce similar structures to those formed during volcano gravitational spreading alone (i.e. grabens, folds, en échelon fractures). Therefore, simple analysis of fault geometry and fault kinetic indicators is not sufficient to distinguish gravitational from intrusive complex spreading, except when the intrusive complex is eccentric from the volcano centre. However, the displacement fields obtained for (1) a solely gravitational spreading volcano and for (2) a gravitational spreading volcano with a growing and spreading intrusive complex are very different. Consequently, deformation fields (like those obtained from geodetic monitoring) can give a strong indication of the presence of a spreading intrusive complex. We compare the models with field observations and geophysical evidence on active volcanoes such as La Réunion Island (Indian Ocean), Ometepe Island (Nicaragua) and eroded volcanic remnants such as Ardnamurchan (Scotland) and suggest that a combination between gravitational and intrusive complex spreading has been active.     

Classification of self-potential anomalies on volcanoes and possible interpretations for their subsurface structure

Self-potential (SP) surveys were conducted on 10 island-arc-type volcanoes in Japan. Large-scale SP anomalies that covered entire volcanic edifices and had simple shapes were observed. Neglecting the local fluctuations, we named the SP anomaly that shows voltage gradients anywhere on the flank a long spatial wavelength (LSW) SP anomaly. We first classified these anomalies into three types: 1) no anomaly (SP profile is flat): 2) “W”-shaped profile: and 3) “V”-shaped profile. In some volcanoes, one flank with a flat SP profile and another with half of the W-shaped SP profile coexist. Next, we compared LSW-SP anomalies with various factors that may be related to SP generation and found that the SP profile is approximately flat on the flank where geothermal manifestations exist. This relationship, which is new aspect of the SP data on volcanoes, is important because it can impose a constraint on the model for SP generation and the subsurface structure. Based on the discussion that SP is generated by an electrokinetic mechanism, we propose that resistivity, zeta potential, and hydraulic radius are the possible parameters that control the existence of LSW-SP anomalies. The new relationship between SP and geothermal manifestation are understood in terms of the presence of hydrothermal system. Although the structure of volcanic edifice is probably spatially complicated, the insight into the SP data presented in this study implies that the flanks with flat SP profiles are relatively weak and have high potential for sector collapse.     

Analogue models of caldera collapse in strike-slip tectonic regimes

Regional-scale faulting, particularly in strike-slip tectonic regimes, is a relatively poorly constrained factor in the formation of caldera volcanoes. To examine interactions between structures associated with regional-tectonic strike-slip deformation and volcano-tectonic caldera subsidence, we made scaled analogue models. Tabular (sill-like) inclusions of creamed honey in a sand/gypsum mix replicated shallow-level granitic magma chambers in the brittle upper crust. Lateral motion of a base plate sited below half the sand/gypsum pack allowed simulation of regional strike-slip deformation. Our experiments modelled: (1) strike-slip deformation of a homogeneous brittle medium; (2) strike-slip deformation of a brittle medium containing a passive magma reservoir; (3) caldera collapse into sill-like magma reservoirs without regional strike-slip deformation; and (4) caldera collapse into sill-like magma reservoirs after regional strike-slip deformation. Our results show that whilst the magma chamber shape principally influences the development and geometry of volcano-tectonic collapse structures, regional-tectonic strike-slip faults (Riedel shears and Y-shears) may affect a caldera’s structural evolution in two main ways. Firstly, regional strike-slip faults above the magma chamber may form a pre-collapse structural grain that is exploited and reactivated during subsidence. Our experiments show that such faults may preferentially reactivate where tangential to the collapse area and coincident with the chamber margins. In this case, volcano-tectonic extension in the caldera periphery tends to localise on regional-tectonic faults that lie just outside the chamber margins. In addition, volcano-tectonic reverse faults may link with and reactivate pre-collapse regional-tectonic faults that lie just inside the chamber margins. Secondly, where regional-tectonic strike-slip faults define corners in the magma chamber margin, they may halt the propagation of volcano-tectonic reverse faults. The experiments also highlight the potential difficulties in assessing the relative contributions of volcano-tectonic and regional-tectonic subsidence processes to the final caldera structure seen in the field. Disruption of the pre-collapse surface by regional-tectonic faulting was preserved during coherent volcano-tectonic subsidence to produce a caldera floor of differentially-subsided fault blocks. Without definitive evidence for syn-eruptive growth faulting, thickness changes in caldera fill across such regional-tectonic fault blocks in nature could be mistaken as evidence for piecemeal volcano-tectonic collapse.     

Dykes and structures of the NE rift of Tenerife, Canary Islands: a record of stabilisation and destabilisation of ocean island rift zones

Many oceanic island rift zones are associated with lateral sector collapses, and several models have been proposed to explain this link. The North–East Rift Zone (NERZ) of Tenerife Island, Spain offers an opportunity to explore this relationship, as three successive collapses are located on both sides of the rift. We have carried out a systematic and detailed mapping campaign on the rift zone, including analysis of about 400 dykes. We recorded dyke morphology, thickness, composition, internal textural features and orientation to provide a catalogue of the characteristics of rift zone dykes. Dykes were intruded along the rift, but also radiate from several nodes along the rift and form en échelon sets along the walls of collapse scars. A striking characteristic of the dykes along the collapse scars is that they dip away from rift or embayment axes and are oblique to the collapse walls. This dyke pattern is consistent with the lateral spreading of the sectors long before the collapse events. The slump sides would create the necessary strike-slip movement to promote en échelon dyke patterns. The spreading flank would probably involve a basal decollement. Lateral flank spreading could have been generated by the intense intrusive activity along the rift but sectorial spreading in turn focused intrusive activity and allowed the development of deep intra-volcanic intrusive complexes. With continued magma supply, spreading caused temporary stabilisation of the rift by reducing slopes and relaxing stress. However, as magmatic intrusion persisted, a critical point was reached, beyond which further intrusion led to large-scale flank failure and sector collapse. During the early stages of growth, the rift could have been influenced by regional stress/strain fields and by pre-existing oceanic structures, but its later and mature development probably depended largely on the local volcanic and magmatic stress/strain fields that are effectively controlled by the rift zone growth, the intrusive complex development, the flank creep, the speed of flank deformation and the associated changes in topography. Using different approaches, a similar rift evolution has been proposed in volcanic oceanic islands elsewhere, showing that this model likely reflects a general and widespread process. This study, however, shows that the idea that dykes orient simply parallel to the rift or to the collapse scar walls is too simple; instead, a dynamic interplay between external factors (e.g. collapse, erosion) and internal forces (e.g. intrusions) is envisaged. This model thus provides a geological framework to understand the evolution of the NERZ and may help to predict developments in similar oceanic volcanoes elsewhere.     

Volcanic edifice stability during cryptodome intrusion

Limit equilibrium analyses were applied to the 1980 Mount St. Helens and 1956 Bezymianny failures in order to examine the influence on stability of structural deformation produced by cryptodome emplacement. Weakening structures associated with the cryptodome include outward-dipping normal faults bounding a summit graben and a flat shear zone at the base of the bulged flank generated by lateral push of the magma. Together with the head of the magmatic body itself, these structures serve directly to localize failure along a critical surface with low stability deep within the interior of the edifice. This critical surface, with the safety coefficient reduced by 25-30%, is then very sensitive to stability condition variation, in particular to the pore-pressure ratio (r_u) and seismicity coefficient (n). For r_u=0.3, or n=0.2, the deep surface suffers catastrophic failure, removing a large volume of the edifice flank. In the case of Mount St. Helens, failure occurred within a material with angle of friction ~40°, cohesion in the range 10⁵-10⁶ Pa, and probably significant water pore pressure. On 18 May 1980, detachment of slide block I occurred along a newly formed rupture surface passing through the crest of the bulge. Although sliding of block I may have been helped by the basal shear zone, significant pore pressure and a triggering earthquake were required (r_u=0.3 and n=0.2). Detachment of the second block was guided by the summit normal fault, the front of the cryptodome, and the basal shear zone. This occurred along a deep critical surface, which was on the verge of failure even before the 18 May 1980 earthquake. The stability of equivalent surfaces at Bezymianny Volcano appears significantly higher. Thus, although magma had already reached the surface, weaker materials, or higher pore pressure and/or seismic conditions were probably required to reach the rupture threshold. From our analysis, we find that deep-seated sector collapses formed by removing the edifice summit cannot generally result from a single slide. Cryptodome-induced deformation does, however, provide a deep potential slip surface. As previously thought, it may assist deep-seated sector collapse because it favors multiple retrogressive slides. This leads to explosive depressurization of the magmatic and hydrothermal systems, which undermines the edifice summit and produces secondary collapses and explosive blasts.     

Back-arc rifting in the Izu-Bonin Island Arc: Structural evolution of Hachijo and Aoga Shima Rifts

Abstract Multi- and single-channel seismic profiles are used to investigate the structural evolution of back-arc rifting in the intra-oceanic Izu-Bonin Arc. Hachijo and Aoga Shima Rifts, located west of the Izu-Bonin frontal arc, are bounded along-strike by structural and volcanic highs west of Kurose Hole, North Aoga Shima Caldera and Myojin Sho arc volcanoes. Zig-zag and curvilinear faults subdivide the rifts longitudinally into an arc margin (AM), inner rift, outer rift and proto-remnant arc margin (PRA). Hachijo Rift is 65 km long and 20–40 km wide. Aoga Shima Rift is 70 km long and up to 45 km wide. Large-offset border fault zones, with convex and concave dip slopes and uplifted rift flanks, occur along the east (AM) side of the Hachijo Rift and along the west (PRA) side of the Aoga Shima Rift. No cross-rift structures are observed at the transfer zone between these two regions; differential strain may be accommodated by interdigitating rift-parallel faults rather than by strike- or oblique-slip faults. In the Aoga Shima Rift, a 12 km long flank uplift, facing the flank uplift of the PRA, extends northeast from beneath the Myojin Knoll Caldera. Fore-arc sedimentary sequences onlap this uplift creating an unconformity that constrains rift onset to ～1-2Ma. Estimates of extension (～3km) and inferred age suggest that these rifts are in the early syn-rift stage of back-arc formation. A two-stage evolution of early back-arc structural evolution is proposed: initially, half-graben form with synthetically faulted, structural rollovers (ramping side of the half-graben) dipping towards zig-zagging large-offset border fault zones. The half-graben asymmetry alternates sides along-strike. The present ‘full-graben’ stage is dominated by rift-parallel hanging wall collapse and by antithetic faulting that concentrates subsidence in an inner rift. Structurally controlled back-arc magmatism occurs within the rift and PRA during both stages. Significant complications to this simple model occur in the Aoga Shima Rift where the east-dipping half-graben dips away from the flank uplift along the PRA. A linear zone of weakness caused by the greater temperatures and crustal thickness along the arc volcanic line controls the initial locus of rifting. Rifts are better developed between the arc edifices; intrusions may be accommodating extensional strain adjacent to the arc volcanoes. Pre-existing structures have little influence on rift evolution; the rifts cut across large structural and volcanic highs west of the North Aoga Shima Caldera and Aoga Shima. Large, rift-elongate volcanic ridges, usually extruded within the most extended inner rift between arc volcanoes, may be the precursors of sea floor spreading. As extension continues, the fissure ridges may become spreading cells and propagate toward the ends of the rifts (adjacent to the arc volcanoes), eventually coalescing with those in adjacent rift basins to form a continuous spreading centre. Analysis of the rift fault patterns suggests an extension direction of N80°E ± 10° that is orthogonal to the trend of the active volcanic arc (N10°W). The zig-zag pattern of border faults may indicate orthorhombic fault formation in response to this extension. Elongation of arc volcanic constructs may also be developed along one set of the possible orthorhombic orientations. Border fault formation may modify the regional stress field locally within the rift basin resulting in the formation of rift-parallel faults and emplacement of rift-parallel volcanic ridges. The border faults dip 45–55° near the surface and the majority of the basin subsidence is accommodated by only a few of these faults. Distinct border fault reflections decreases dips to only 30° at 2.5 km below the sea floor (possibly flattening to near horizontal at 2.8 km although the overlying rollover geometry shows a deeper detachment) suggesting that these rifting structures may be detached at extremely shallow crustal levels.     

A gravitational spreading origin for the Socompa debris avalanche

Socompa Volcano arguably provides the world's best-exposed example of a sector collapse-derived debris avalanche deposit. New observations lead us to re-interpret the origin of the sector collapse. We show that it was triggered by failure of active thrust-anticlines in sediments and ignimbrites underlying the volcano. The thrust-anticlines were a result of gravitational spreading of substrata under the volcano load. About 80% of the resulting avalanche deposit is composed of substrata formerly residing under the volcano and in the anticlines. The collapse scar can be traced up to 5 km from the edifice, truncating two spreading-related anticlines, which collapsed in the event. Outcrops near the volcano preserve evidence of edifice material being carried along on top of mobilised substrata. On the north side of the scar, the avalanche motion was initially at right angles to the failure edge. Structural relations indicate that immediately prior to collapse the substrata disintegrated, became effectively liquidised, and were ejected from beneath the edifice. Catastrophic mobilisation of substrata probably resulted from breakdown of ignimbrite clasts and cement. It may have occurred through progressive rock fracture by high shear strain during spreading. Material ejected from under Socompa formed a layer on which volcanic edifice debris was transported. This interpretation of events explains the puzzling observation that avalanche units with the lowest gravitational potential energy moved the furthest. It can also account for avalanche motion normal to the collapse scar walls. Ignimbrites and other rock types probably capable of similar behaviour underlie many other volcanoes. Identification of spreading at other sites could therefore be a first step towards assessment of the potential for this style of catastrophic sector collapse.     

Debris avalanches and debris flows transformed from collapses in the Trans-Mexican Volcanic Belt, Mexico – behavior, and implications for hazard assessment

Volcanoes of the Trans-Mexican Volcanic Belt (TMVB) have yielded numerous sector and flank collapses during Pleistocene and Holocene times. Sector collapses associated with magmatic activity have yielded debris avalanches with generally limited runout extent (e.g. Popocatépetl, Jocotitlán, and Colima volcanoes). In contrast, flank collapses (smaller failures not involving the volcano summit), both associated and unassociated with magmatic activity and correlating with intense hydrothermal alteration in ice-capped volcanoes, commonly have yielded highly mobile cohesive debris flows (e.g. Pico de Orizaba and Nevado de Toluca volcanoes). Collapse orientation in the TMVB is preferentially to the south and northeast, probably reflecting the tectonic regime of active E–W and NNW faults. The differing mobilities of the flows transformed from collapses have important implications for hazard assessment. Both sector and flank collapse can yield highly mobile debris flows, but this transformation is more common in the cases of the smaller failures. High mobility is related to factors such as water content and clay content of the failed material, the paleotopography, and the extent of entrainment of sediment during flow (bulking). The ratio of fall height to runout distance commonly used for hazard zonation of debris avalanches is not valid for debris flows, which are more effectively modeled with the relation inundated area to failure or flow volume coupled with the topography of the inundated area.     

A model of diffuse degassing at three subduction-related volcanoes

2 and δ¹³C in soil gas were measured at three active subduction-related stratovolcanoes (Arenal and Poás, Costa Rica; Galeras, Colombia). In general, Rn, CO₂ and δ¹³C values are higher on the lower flanks of the volcanoes, except near fumaroles in the active craters. The upper flanks of these volcanoes have low Rn concentrations and light δ¹³C values. These observations suggest that diffuse degassing of magmatic gas on the upper flanks of these volcanoes is negligible and that more magmatic degassing occurs on the lower flanks where major faults and greater fracturing in the older lavas can channel magmatic gases to the surface. These results are in contrast to findings for Mount Etna where a broad halo of magmatic CO₂ has been postulated to exist over much of the edifice. Differences in radon levels among the three volcanoes studied here may result from differences in age, the degree of fracturing and faulting, regional structures or the level of hydrothermal activity. Volcanoes, such as those studied here, act as plugs in the continental crust, focusing magmatic degassing towards crater fumaroles, faults and the fractured lower flanks. Received: 16 December 1997 / Accepted: 27 January 2000     

The role of the Pernicana Fault System in the spreading of Mt. Etna (Italy) during the 2002–2003 eruption

Flank instability and collapse are observed at many volcanoes. Among these, Mt. Etna is characterized by the spreading of its eastern and southern flanks. The eastern spreading area is bordered to the north by the E–W-trending Pernicana Fault System (PFS). During the 2002–2003 Etna eruption, ground fracturing along the PFS migrated eastward from the NE Rift, to as far as the 18 km distant coastline. The deformation consisted of dextral en-echelon segments, with sinistral and normal kinematics. Both of these components of displacement were one order of magnitude larger (~1 m) in the western, previously known, portion of the PFS with respect to the newly surveyed (~9 km long) eastern section (~0.1 m). This eastern section is located along a pre-existing, but previously unknown, fault, where displaced man-made structures give overall slip rates (1–1.9 cm/year), only slightly lower than those calculated for the western portion (1.4–2.3 cm/year). After an initial rapid motion during the first days of the 2002–2003 eruption, movement of the western portion of the PFS decreased dramatically, while parts of the eastern portion continued to move. These data suggest a model of spreading of the eastern flank of Etna along the PFS, characterized by eruptions along the NE Rift, instantaneous, short-lived, meter-scale displacements along the western PFS and more long-lived centimeter-scale displacements along the eastern PFS. The surface deformation then migrated southwards, reactivating, one after the other, the NNW–SSE-trending Timpe and Trecastagni faults, with displacements of ~0.1 and ~0.04 m, respectively. These structures, along with the PFS, mark the boundaries of two adjacent blocks, moving at different times and rates. The new extent of the PFS and previous activity over its full length indicate that the sliding eastern flank extends well below the Ionian Sea. The clustering of seismic activity above 4 km b.s.l. during the eruption suggests a deep décollement for the moving mass. The collected data thus suggests a significant movement (volume >1,100 km³) of the eastern flank of Etna, both on-shore and off-shore.Editorial responsibility: R. Cioni     

Volcano-tectonic controls and emplacement kinematics of the Iriga debris avalanches (Philippines)

Mt Iriga in southeastern Luzon is known for its spectacular collapse scar that was possibly created in 1628 ad by a 1.5-km³ debris avalanche. The debris avalanche deposit (DAD) covered 70?km² and dammed the Barit River to form Lake Buhi. The collapse has been ascribed to a non-volcanic trigger related to a major strike-slip fault under the volcano. Using a combination of fieldwork and remote sensing, we have identified a similar size, older DAD to the southwest of the edifice that originated from a sector oblique to the underlying strike-slip fault. Both deposits cover wide areas of low, waterlogged plains, to a distance of about 16?km for the oldest and 12?km for the youngest. Hundreds of metre-wide and up to 50-m-high hummocks of intact conglomerate, sand and clay units derived from the base of the volcano show that the initial failure planes cut deep into the substrata. In addition, large proportions of both DAD consist of ring-plain sediments that were incorporated by soft-sediment bulking and extensive bulldozing. An ignimbrite unit incorporated into the younger DAD forms small (less than 5?m high) discrete hummocks between the larger ones. Both debris avalanches slid over water-saturated soft sediment or ignimbrite and spread out on a basal shear zone, accommodated by horst and graben formation and strike-slip faults in the main mass. The faults are listric and flatten into a well-developed basal shear zone. This shear zone contains components from the substrate and has a diffuse contact with the intact substrata. Long, transport-normal ridges in the distal parts are evidence of compression related to deceleration and bulldozing. The collapse orientation and structure on both sectors and DAD constituents are consistent with collapse being a response to combined transtensional faulting and gravity spreading. Iriga can serve as a model for other volcanoes, such as Mayon, that stand in sedimentary basins undergoing transtensional strike-slip faulting.     

Light hydrocarbons in hydrothermal and magmatic fumaroles: hints of catalytic and thermal reactions

Volcanic gaseous mixtures emitted from active volcanoes frequently show variable amounts of saturated (alkanes), unsaturated (alkenes) and aromatic volatile hydrocarbons. Three major patterns of distributions can be recognized, apparently related to the chemical-physical environment of formation of the gas exhalations: alkane-rich, low-temperature gas emissions from recently active volcanic areas; aromatic-rich hydrothermal manifestations; and alkene-rich, magmatic fumaroles on active volcanoes. Thermodynamic data, together with theoretical and practical findings from the petroleum industry, point to two main types of reactions occurring in these volcanic environments: cracking and reforming. Cracking processes, mainly caused by thermal effects, occur when hydrocarbon-bearing hydrothermal fluids enter and mix with a hot and dry, rapidly rising magmatic gas phase. The most probable products are light alkenes with carbon numbers decreasing with increasing reaction temperatures. The presence of aromatic species in hydrothermal fluids can be linked to reforming processes, catalysed by several possible agents, such as smectites and zeolites, generally present in the hydrothermally altered volcanic terranes, and facilitated by long residence times in a hydrothermal envelope.