Rock Avalanche Deposits Research Articles

Rock avalanches that travel onto glaciers and related developments, Karakoram Himalaya, Inner Asia

Catastrophic rock slope failures in rugged terrain commonly generate rock avalanches. When these occur in glacier basins the extent of landslide run out and its emplacement geometry are affected by movement over ice. Substantial modifications of glacier activity and sedimentation can follow. Ice activity leads to rapid modification, transport and dispersal of landslide debris. The nature of these interactions is described from observations over 20 years at Bualtar Glacier in the Karakoram. In August 1986 rock avalanches descended onto the ablation zone of the glacier. Their moisture content and patterns of deposition were affected by travel over ice. Glacier movement increased sharply at and below the rock avalanche deposits and, within a few months, the glacier surged. A second surge occurred 2 years later. Major slope failures and debris flows were triggered beyond the ice margins, and ponded melt water led to small outburst floods. By 2005 the landslide material had been transported some 9 km, about one third of this distance in the surges. Eventually it was fully reworked to become less readily distinguishable from other heavy supraglacial debris. A large area of thickened ice persisted where the debris reduced ablation; a positive impact on mass balance equivalent to a 20% increase in annual accumulation. However, it occurred as a moving segment of the ablation zone. In 10 years it had, in effect, replaced the mass transferred in the surges. Data for the mid-1980s indicate the rock avalanches exceeded pre-existing supraglacial debris by roughly five times and, over a 30 year period, will equal almost 500 years of normal supraglacial transport to the glacier margins. Other impacts of the episode suggest a doubling of this contribution to denudation. Similar developments were observed at three other glaciers where recent rock avalanches occurred and at twenty-one prehistoric rock avalanches newly identified as having spread over glaciers. It seems inadequate to treat such events only as singular extremes, or the landslide and glacial processes separately. Here they are interpreted as contributing to ‘disturbance regime’ landscapes, whose history reflects episodic and regionally specific responses to strong interference among different earth surface processes. They generate polygenetic deposits and landforms that are morphologically complex. Individual episodes involve decades to millennia; their numbers and repetition can over-ride or redirect geotectonic and climatic conditions. It is suggested they are major, but hitherto misread or neglected, elements of landform evolution in the Karakoram and, perhaps, in other high mountains.

Characterization of a rock avalanche deposit for risk assessment in the town of Celano (Fucino Basin, Central Italy)

Abstract. This paper describes multidisciplinary investigations carried out in the urban centre of Celano, a small town located at the northern edge of the Fucino Basin (Central Italy). The town lies upon a wide debris body that was recognized in this study as a rock avalanche deposit estimated to date to the Holocene. Geomorphologic studies and geophysical investigations led to a detailed characterization of the landslide deposit and the surrounding units. The information obtained was used to assess the vulnerability of the Celano municipal area, by evaluating the stability of the landslide body and the behaviour of its lithologies under seismic loads.

Stateline fault system: A new component of the Miocene-Quaternary Eastern California shear zone

The Eastern California shear zone is an active, north-northwest–trending zone of intraplate right-lateral shear that absorbs ∼25% of Pacific-North America relative plate motion. The Stateline fault system (SFS), which includes several previously recognized, discontinuously exposed Quaternary structures along the California-Nevada border, is in this paper defined as a continuous, 200-km–long zone of active dextral shear that includes (from south to north) the Mesquite, Pahrump, and Amargosa Valley segments. Recognition of this system expands the known extent of the Eastern California shear zone ∼50 km to the east-northeast from its traditionally recognized boundary along the Death Valley fault system. Proximal volcanic and rock avalanche deposits offset across the Mesquite segment of the SFS indicate 30 ± 4 km of slip on this structure since 13.1 ± 0.2 Ma. This offset is an order of magnitude larger than previous estimates across this section of the SFS, but it is consistent with larger offsets previously proposed for the central and northern sections. The total offset and averaged slip rate since mid-Miocene time (2.3 ± 0.35 mm/yr) are similar to those of other major faults across this portion of the Basin and Range, which, from east to west, include the Death Valley, Panamint Valley-Hunter Mountain, and Owens Valley fault systems. However, in contrast to these faults, the average post–mid-Miocene slip rate on the SFS is approximately twice that estimated from present-day geodetic observations and an order of magnitude greater than estimates of average post–mid-Pleistocene slip rates. This discrepancy between long-term, short-term, and geodetically derived slip rates differs from other geologic-geodetic, slip-rate discrepancies in the Eastern California shear zone, where geodetic slip rates are significantly faster than both long-term and short-term geologic slip rates. This suggests that either the slip rate on the SFS has diminished over time, such that the system is an abandoned strand of the relatively young Eastern California shear zone, or that the present-day slip rate represents a transient period of slow slip, such that strands of the shear zone must accommodate a complex spatial and temporal distribution of slip.

Chronology of Holocene rock-avalanche deposits based on Schmidt-hammer relative dating and dust stratigraphy in nearby bog deposits, Vora, inner Nordfjord, Norway

A number of large rock avalanches from the northern slope of the 1435 m high mountain of Vora in Breim, inner Nordfjord have been mapped. The avalanche tongues dam the 4 km long Sandalsvatnet, they cover an area of 3 km 2 and the volume exceeds 100 million m3. Age estimates have been obtained by radiocarbon datings of rock dust layers in a bog and the Schmidt-hammer method. Altogether 11 large rock avalanches have occurred throughout the Holocene, of which six lobes can be recognized morphologically. Five silt laminae in a 5.5 m deep bog basin distally to the avalanche lobes are interpreted as primary sediments from dust clouds generated by rock avalanches. In addition three silt beds are interpreted as glaciolacustrine sediments related to temporary damming caused by landsliding. Average avalanche frequency during the Holocene was c. 1 per 1000 yr and during the period 7100—3600 BP it was c. 1 per 700 yr. After 3600 BP no rock avalanches have been recognized.

10Be exposure ages of a rock avalanche and a late glacial moraine in Alta Valtellina, Italian Alps

A rock avalanche deposit was investigated in order to understand the chronological evolution of geological hazards and to evaluate the interaction of the triggering geodynamic processes in the valley Val Viola, Italian Alps. The deposit is situated west of the Alpe Dosdé, in a permafrost area with deep-seated gravitational deformations (DSGD) along a tectonic line. Based on its geomorphologic context, the rock avalanche was first interpreted as a result of slope stress release without exact timing. This hypothesis was tested by measuring the 10Be exposure date of quartz from one boulder from the rock avalanche. The age of 7430±460 years places the event in the early Holocene. The timing of the last deglaciation was constrained using the inner late glacial moraine of a moraine doublet in the valley Alpe Dosdé situated at an altitude between 2140 and 2120 m a.s.l. west of the rock avalanche. The 10Be concentrations of quartz yield minimum exposure ages of 11,480±670 and 10,850±820 years. Different proposals for potential triggering factors of the rock avalanche include (a) melting of the local valley glacier and slope stress release in the Val Viola, likely to play a minor role as trigger, because of the time delay between the deglaciation and the rock avalanche event. More likely are (b) enhanced crustal seismicity induced by post-glacial regional isostatic glacial rebound coupled with tectonic stress or/and (c) climate conditions with higher temperatures around 7430±460 years, resulting in an upwards movement of the permafrost limit and destabilization of the rock walls.

The Keylong Serai rock avalanche, NW Indian Himalaya: geomorphology and palaeoseismic implications

This paper describes the geomorphology of rock avalanche deposits that resulted from a major mountain slope failure at Keylong Serai on the north slope of the Indian High Himalaya, an area of high altitude desert. Cosmogenic 10Be exposure ages of the widespread deposits indicate their formation 7,510 ± 110 years BP. Proxy records for this region of the Himalaya imply a similar dry climatic regime to the present day at this time, suggesting that precipitation was an unlikely trigger for this rock avalanche. An alternative mechanism associated with rock-wall stress relaxation is also unlikely, given the earlier timing of deglaciation in this area. Given the enormous volume of debris generated by this event, the most likely trigger for this mountain collapse and resultant rock avalanche is high ground acceleration during a great earthquake (M > 8). It is proposed that rock avalanches can be used to extend the limited palaeoseismic record and improve information on the recurrence interval of great earthquakes within the Himalaya arc.

The Poverty Hills, Owens Valley, California--Transpressional Uplift or Ancient Landslide Deposit?

Research Article| April 01, 2006 The Poverty Hills, Owens Valley, California—Transpressional Uplift or Ancient Landslide Deposit? KIM M. BISHOP; KIM M. BISHOP 1Department of Geology, California State University, Los Angeles, 5151 State University Drive, Los Angeles, CA 90032 Search for other works by this author on: GSW Google Scholar STEVE CLEMENTS STEVE CLEMENTS 2SCS Engineers, 6601 Koll Center Parkway, Suite 140, Pleasanton, CA 64566 Search for other works by this author on: GSW Google Scholar Author and Article Information KIM M. BISHOP 1Department of Geology, California State University, Los Angeles, 5151 State University Drive, Los Angeles, CA 90032 STEVE CLEMENTS 2SCS Engineers, 6601 Koll Center Parkway, Suite 140, Pleasanton, CA 64566 Publisher: Association of Environmental & Engineering Geologists First Online: 09 Mar 2017 Online ISSN: 1558-9161 Print ISSN: 1078-7275 © 2006 Association of Engineering Geologists Environmental & Engineering Geoscience (2006) 12 (4): 301–314. https://doi.org/10.2113/gseegeosci.12.4.301 Article history First Online: 09 Mar 2017 Cite View This Citation Add to Citation Manager Share Icon Share Facebook Twitter LinkedIn MailTo Tools Icon Tools Get Permissions Search Site Citation KIM M. BISHOP, STEVE CLEMENTS; The Poverty Hills, Owens Valley, California—Transpressional Uplift or Ancient Landslide Deposit?. Environmental & Engineering Geoscience 2006;; 12 (4): 301–314. doi: https://doi.org/10.2113/gseegeosci.12.4.301 Download citation file: Ris (Zotero) Refmanager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu toolbar search search input Search input auto suggest filter your search All ContentBy SocietyEnvironmental & Engineering Geoscience Search Advanced Search Abstract The Poverty Hills are a 5 km2, 300-m-high mass of plutonic and meta-sedimentary rocks exposed along the axis of the alluviated Owens Valley graben in eastern California. Two models have been proposed for the origin of the hills. One model suggests the hills represent a transpressional uplift along a 3-km left-step in the right-lateral Owens Valley fault zone, and the other proposes that the hills are a long-runout rock-avalanche deposit. This article argues that the rock-avalanche hypothesis best fits the geology. The transpressional model is problematic because the fault segment to which right-lateral displacement is purportedly transferred has been shown to have pure dip-slip displacement. The rock-avalanche model is based primarily on lithologic evidence. Throughout the hills are scattered outcrops displaying mosaic, jigsaw, and crackle breccia textures, all of which are characteristic of rock-avalanche deposits. Also, well-exposed roadcut outcrops display preserved source-rock bedding within the matrix-poor breccia framework. Though preserved, the bedding has been contorted and distorted by cataclastic flow of the breccia, another common characteristic of rock-avalanche deposits created during emplacement. The landslide is younger than 3 Ma, the age of inception of the Owens Valley graben, and older than 640 ± 50 ka, the age of a basalt flow that post-dates the landslide. An area in the Inyo Mountains southeast of the hills appears to be the most likely source. Interpretation of the Poverty Hills as a landslide mass suggests an alternative model for the Owens Valley fault-zone kinematics. You do not have access to this content, please speak to your institutional administrator if you feel you should have access.

Disturbance regime landscapes: mountain drainage systems interrupted by large rockslides

The paper examines the role of rockslide-rock avalanches in mountain landscapes, and the landforms associated with them. While the landslides are extremely short-lived events, rock wall detachment scars and rock avalanche deposits can persist for long periods as influences on landscape development. Especially significant are rock avalanches with complex runout and emplacement related to interactions with rugged terrain or deformable substrates. Their characteristics greatly increase the scope of landscape disturbance. Hundreds of rock avalanches are now known, worldwide, that have formed crossvalley barriers interrupting mountain drainage systems. Many have done so for millennia or tens of millennia. They give rise to distinctive sediment assemblages, constructional and erosion landforms generated by other processes responding to the landslides and constrained by them. A landslide interruption epicycle of five phases is described, and related sediment assemblages. These provide the basis for defining a landslide interrupted valley landsystem. Its full significance is seen in mountain drainage basins affected by multiple landslide interruptions. These create naturally fragmented fluvial systems, in which a disturbance regime geomorphology is identified. Stream profiles, sediment delivery, and related landforms are kept in a chronic state of disequilibrium with respect to climatic and geotectonic controls, and drainage organization. The transHimalayan Upper Indus Basin provides an example, a large high mountain drainage system fragmented by more than 170 late Quaternary rock avalanches. In this case, as elsewhere, misidentification of rock avalanches led to neglect of their role in Quaternary histories. The nature and limitations of disturbance regime geomorphology are discussed, and broader implications for mountain landscapes.

The Acheron rock avalanche, Canterbury, New Zealand—morphology and dynamics

The 1100-year-old Acheron rock avalanche deposit lies in an active tectonic setting in Canterbury, New Zealand, and has a volume of ten million cubic metres and a runout distance 3.5 km. The deposit comprises intensely fragmented greywacke rock, and the processes of intense rock fragmentation during runout are postulated to have generated an isotropic dispersive stress. Dynamic simulation shows that the runout can be explained as a flow of dry granular material with a normal coefficient of friction, if the presence of an isotropic dispersive stress within the moving rock debris throughout the runout is assumed. The dispersive stress distribution required to model the rock avalanche runout and match velocities calculated from run-up traces is closely similar to that used to simulate the runout of the much larger Falling Mountain rock avalanche in a similar lithologic and tectonic setting. Both events thus behaved in a fundamentally similar fashion.

Large-scale gravity sliding in the Miocene Shadow Valley Supradetachment Basin, Eastern Mojave Desert, California

The Miocene Shadow Valley basin in the eastern Mojave Desert of California developed above the active west-dipping Kingston Range-Halloran Hills extensional detachment fault system between 13.5 and ca. 7 mybp. Although mass-wasting processes are common phenomena in supradetachment basins, the Shadow Valley basin is an exceptional locale for the study of such processes, especially rock-avalanches and gravity sliding. A score of megabreccias, interpreted as rock-avalanche deposits, and half that number of very large (> 1 km 2, up to 200 m thick), internally intact gravity-driven slide sheets are interbedded with various sedimentary facies. The slide sheets, variably composed of Proterozoic crystalline rocks and Proterozoic, Paleozoic, and Tertiary sedimentary strata, moved across both depositional and erosional surfaces in the basin. Although the majority consist of Paleozoic carbonate rocks, the largest slide sheet, the Eastern Star crystalline allochthon, contains Proterozoic gneisses and their sedimentary cover and is now preserved as klippen atop Miocene lacustrine and alluvial fan deposits over an area > 40 km 2. Estimates of slide sheet runouts into the basin from higher eastern and northern source terranes range from approximately a few km to > 10 km; in most cases the exact provenances of the slide blocks are not known. The basal contacts of Shadow Valley slide sheets are characteristically knife sharp, show few signs of lithologic mixing of upper- and lower-plate rocks, and locally exhibit slickensided and striated, planar fault-like bases. Pronounced folding of overridden Miocene lacustrine and fan deposits beneath the Eastern Star allochthon extends to depths up to 40 m at widely scattered localities. We conclude that this slow moving slide sheet encountered isolated topographic asperities (hills) and that stress transfer across the basal slide surface produced folding of footwall strata. Synkinematic gypsum veins in footwall playa sediments, with fibers up to 12 cm long, have trends and shear senses compatible with the direction and sense of displacement of the overriding crystalline allochthon. The undisturbed veins, which closely parallel the base of the slide sheet, attest to high fluid presence and pressure in the playa sediments—factors facilitating allochthon movement across them. The long length of the fibers, indicative of a protracted dilational process, is incompatible with a catastrophic rate of emplacement. We believe that the only explanation for slow displacement of this allochthon and other gravity driven slide sheets across the landscape is that they formed as slumps on high, steep bedrock slopes and that their elevated heads drove their toes across lower fan and playa deposits. Initial detachments from bedrock sources were facilitated by pre-existing structural and stratigraphic anisotropies. Detachment of the Eastern Star allochthon from the bedrock of Shadow Mountain likely occurred by inversion along the playa-ward dip of a preexisting Mesozoic thrust fault within Proterozoic rock units.

Distinguishing between primary and secondary emplacement events of blocky volcanic deposits using rock size distributions

Rock size characteristics in rock avalanches and blocky lava flows are different and statistical techniques can be used both to distinguish between and to constrain the different emplacement mechanisms of these two processes. At the Chaos Jumbles rock avalanche deposit in California, rock sizes decrease as a function of distance. The rock avalanche mechanism is an example of a secondary process that can only break existing blocks. In contrast, lava flows at Sabancaya, Peru, were emplaced by primary mechanisms capable of creating as well as breaking blocks. These deposits exhibit uniform block sizes as functions of distance and significantly larger blocks overall than Chaos Jumbles. Rock sizes at Inyo domes, also emplaced by a primary process, are significantly smaller in areas with compressional ridges as compared to vent or jumbled areas.

A large landslide on the urban fringe of metropolitan Phoenix, Arizona

A granitic rock avalanche, one of the largest Quaternary landslides in Arizona outside the Grand Canyon with a volume of approximately 5.25 M m 3 and a width a little under 0.5 km, ran ∼1 km from the eastern McDowell Mountains. With lateral levees and pressure ridges, the rock avalanche deposit displays many features found on classic sturzstroms. Failure occurred along a major joint plane paralleling the slope with a dip of 44°, when a major base level lowering event in the Salt River system would have undermined the base of the failed slope, and probably during a period of more moisture than normally available in the present-day arid climate. Failure at the subsurface weathering front highlights the importance of the dramatic permeability change between grussified regolith and relatively fresh bedrock. Rock varnish microlaminations (VMLs) dating, in concert with other geomorphic evidence, suggests that the rock avalanche deposit is slightly older than ∼500 ka. The rock vanish results also have important implications for sampling strategies designed to use cosmogenic nuclide to date Quaternary landslide deposits. Discovery of a large landslide in close proximity to the extending urban fringe of metropolitan Phoenix argues for a more careful analysis of landslide hazards in the region, especially where rapid development excavates bedrock at the base of steep mountain slopes and where the subsurface weathering front is near the surface.

A rock avalanche deposit in Burtness Comb, Lake District, northwest England

AbstractA large tongue‐shaped accumulation of coarse rock debris in Burtness Comb, Lake District, northwest England, previously regarded as a glacial moraine or relict rock glacier, has been examined and is reinterpreted as the product of rock‐slope failure. The debris tongue has morphological similarities with landforms produced by rock avalanches. This is the first rock avalanche deposit to be reported from the Lake District. Schmidt hammer R‐values and morphological contrasts between the upper and lower parts of the debris imply that two stages of slope failure and downslope debris streaming occurred. In the absence of evidence for a Loch Lomond Stade (c. 12.9–11.5 cal. ka BP) glacier in Burtness Comb, the debris may have accumulated at any time since removal of the Dimlington Stade ice cover (c. 16.8 cal. ka BP) and is regarded as a form of paraglacial landscape response to the effects of glaciation. Copyright © 2004 John Wiley & Sons, Ltd.

Rock avalanching into a landslide-dammed lake causing multiple dam failure in Las Conchas valley (NW Argentina) ? evidence from surface exposure dating and stratigraphic analyses

Generally landslide dams which exist for several hundreds to thousands of years are considered as stable. We show with an example from the Argentine Andes that such dams can exist for several thousands of years but still may fail catastrophically. Multiple rock avalanches impounded two lakes with surface areas of ~8 km2 and ~600 km2, respectively, in Las Conchas valley, NW Argentina. Surface exposure dating (SED) by 10Be of the rock-avalanche deposits or landslide scars indicates that these landslides occurred at 15,300±2,000 yr and 13,550±900 yr. The dams were stable during a strong earthquake, as suggested by seismites within related lake sediments and by multiple coeval landslides in this region, which occurred at ~7.5 kyr. However, when a further rock-avalanche fell into the lower, smaller lake at 4,800±500 yr the dam downriver was destroyed, presumably by the resulting tsunami wave. The resulting flood also destroyed an additional rock-fall dam which had formed at ~5,630 yr 14C cal BP 30 km downriver. The new dam formed by the second rock avalanche was eroded prior to 3,630 yr 14C cal BP. This dam erosion coincides with an important climatic shift towards more humid conditions in the Central Andes. Our results show that instead of direct effects of strong seismicity on landslide dams, (1) landsliding into a landslide-dammed lake, (2) abrupt hydrological changes, and (3) climate change towards conditions related to enhanced run-off are processes which can produce failures of “quasi-stable” natural dams.

The Beinn Alligin rock avalanche, NW Scotland: cosmogenic 10Be dating, interpretation and significance

A tongue of very coarse rockslide debris that extends 1.25km downvalley below Beinn Alligin in NW Scotland has been variously interpreted as a glacier-cored rock glacier, landslide debris redistributed by glacier ice or an excess-runout landslide. Exposure dating with cosmogenic 10Be demonstrates that the the debris mass was emplaced at 3950±320 yr BP, and therefore was not associated with glacier ice. Calculations based on frictional considerations imply that the feature is an excess-runout rock avalanche (sturzstrom) deposit. The morphological characteristics of the deposit appear consistent with movement by grainflow or fragmental flow. Failure is inferred to reflect time-dependent paraglacial stress release and consequent propagation of an internal joint network, but may have been triggered by seismic activity. The late-Holocene age of failure implies persistence of the eŒects of paraglacial stress release over a time-scale of several millennia.

Entrainment of debris in rock avalanches: An analysis of a long run-out mechanism

Many rock avalanches entrain and liquefy saturated soil from their paths. Evidence for this includes mud displaced from the margins of rock avalanche deposits, substrate material smeared along the base of deposits, extrusion of liquefied soil upward through the deposits, and increases of volume. A hypothesis first suggested in 1881 and since reinforced by several authors suggests that entrainment of substrate material increases mobility. Although the process has been discussed in the literature for more than 100 years, few detailed and quantitative descriptions exist. The main purpose of this paper is to describe two recent cases from British Columbia, Canada, where rockslides entrained substrate on a very large scale, influencing the character of the events. Estimated volume balance curves, based on detailed field mapping, are provided for both cases. Dynamic analyses are carried out using a numerical model and using the same set of rheological parameters. The mechanism of material entrainment and displacement is discussed. The data suggest that rapid rock failures entraining very large quantities of saturated substrate material represent a special type of landslide, transitional between rock avalanche and debris avalanche. Many rock avalanches can thus be seen as end members of a continuum of phenomena involving rock failure followed by interaction with saturated substrate.

Discrimination of hot versus cold avalanche deposits: Implications for hazard assessment at Mount Meager, B.C.

Abstract. The surficial deposits surrounding the Mount Meager volcanic complex include numerous avalanche deposits. These deposits share many attributes: (a) they are nearly monolithologic and comprise mainly intermediate volcanic rock clasts, (b) they lack internal structure, and (c) they are very poorly sorted. Despite these similarities, the avalanche deposits represent two distinct processes. Mass wasting of the Mount Meager volcanic edifice has produced cold rock avalanche deposits, whereas gravitational collapse of active lava domes and flows has produced hot block and ash avalanche deposits. The ability to discriminate between these "hot" and "cold" avalanche deposits is a critical component in the assessment of hazards in volcanic terranes. Hot block and ash avalanche deposits can be distinguished by the presence of radially-oriented joints, breadcrust textures, and incipient welding, which are features indicative of high emplacement temperatures. Conversely, rock avalanche deposits resulting from mass wasting events may be distinguished by the presence of clasts that preserve pre-depositional weathering and jointing surfaces. Volcanic avalanches are mechanically similar to rock avalanches but pose a greater hazard due to high temperatures, increased fluidization from degassing and the potential to decouple highly mobile elutriated ash clouds. The increasing use of hazardous regions such as the Lillooet River valley requires more reliable risk assessment in order to minimize losses from future hazardous events.

Granular memory and its effect on the triggering and distribution of rock avalanche events

Rock avalanches are uncommon, large, catastrophic events in which granular masses of rock flow at high speeds, commonly with unusually long runouts. Necessary conditions for rock avalanche occurrence include high relief, steep slopes, and a prefractured (granular) groundmass. An additional parameter, antecedent shear of the granular mass in the direction of failure, may be prerequisite. In many instances the basement mass of the eventual landslide underwent internal strain subparallel to the direction of mass wasting, commonly focused along a basal surface (e.g., fault). New experimental data from laboratory experiments indicate that shear of a granular mass produces subtle anisotropy in grain arrangements that strongly affects the subsequent distribution of deformation and rearrangement, the failure threshold, and the rate at which steady state conditions are reached. This phenomenon may help to explain the spatial clustering of rock avalanche deposits. If memory of prior shear is also relevant on larger scales, taking deformation history into account could greatly enhance prediction of rock avalanche occurrence.

Geological history and potential hazards of the late-Pleistocene to Recent Plat Pays volcanic complex, Dominica, Lesser Antilles

During 1998–2000, the island of Dominica in the Lesser Antilles experienced a major volcanic earthquake swarm spatially associated with volcanic centres in the south of the island. This swarm provided the motivation for a major re-assessment of geologic history and volcanic hazards in this region. This has led to a re-interpretation of the south-western-most corner of the island as a large volcanic complex, termed the Plat Pays volcanic complex, which has exhibited a wide range of past eruptive activity and which, in the Quaternary, experienced a major caldera collapse triggered by a voluminous pyroclastic eruption. Stratigraphy and new 14C age determinations reveal a complicated history of development. The earliest activity is represented by the formation of the Plat Pays stratovolcano and associated parasitic domes. A major explosive eruption ~ 39,000 years B.P. produced the Grand Bay ignimbrite and triggered a major caldera collapse of the summit and south-western flank of the Plat Pays stratovolcano. Following the major eruption, re-injection of Plat Pays magma resulted in the extrusion of approximately 12 lava domes both within and outside the resulting depression. Caldera collapse was followed by at least one catastrophic flank collapse, but it is unclear whether or not this was triggered by caldera formation. The only on-shore evidence of flank collapse is the breach of the caldera rim and truncation of post-caldera deposits emplaced on or near the caldera rim; we find no evidence of on-shore large-magnitude rock avalanche deposits within the stratigraphic framework of the Plat Pays volcanic complex. Frequent seismic swarms of volcanic earthquakes and vigorous geothermal activity indicate that south Dominica is still underlain by an active magma reservoir system. Our new geological observations, combined with an interpretation of recent seismic swarm activity, suggest that an eruption (probably dome-forming) from the Plat Pays volcanic complex is probable within the next 100 years.

Dynamic simulation of the motion of fragmenting rock avalanches

A mass-referenced continuum model for dynamic analysis of rapid mass movement (DAN) is verified by laboratory and field data. Increased earth pressure coefficients are used in this model to represent the dispersive pressures caused by fragmentation within a translating rock avalanche. The numerical model demonstrates that increased runout in large rock avalanches can occur with normal friction coefficients if higher than normal internal pressures, such as those believed to be generated by fragmentation, are present. The extent of the Falling Mountain rock-avalanche deposit in New Zealand is reproduced in the model with normal friction and high earth pressure coefficients to represent by analogy the additional internal pressures due to fragmentation. It appears that if internal friction is changed by fragmentation, it is only by a small amount and may increase rather than decrease. To test this, and to move beyond the present analogue model, requires a better understanding of the rheology of fragmenting rock.Key words: rock avalanches, long runout, fragmentation, simulation model, dispersive stresses, earth pressure coefficients, Falling Mountain.