Volcanoes are dynamic evolving structures, with life cycles that are punctuated by episodes of flank instability. Putauaki (Mount Edgecumbe) is a stratovolcano located onshore in the Bay of Plenty, New Zealand. The aim of this study was to assess the stability of Putauaki and analyse the risk associated with volcanic collapse. To achieve this objective, a multidisciplinary approach was used, incorporating geomorphological and geological mapping, rock mass classification, laboratory testing to identify geotechnical properties of materials representative of the volcano, stability modelling, and analysis of landslide run-out zones. Putauaki comprises two predominant features including the larger and younger Main Cone (the summit lying 820 m a.s.l., slope angles up to 36 ), and smaller and older Main Dome (the summit lying 420 m a.s.l., slope angle of 24 ). Both features show little evidence of erosion or surface water. Rock mass description defined six lithotechnical units including indurated andesite, indurated dacite, scoriaceous andesite, altered andesite (all categorised as hard rocks), and block and ash flow and Matahina Ignimbrite (both categorised as soft rocks). The uniaxial compressive strength (UCS) of indurated andesite and indurated dacite was 60 4 MPa and 44.7 0.9 MPa respectively, correlating with moderately strong rock. Discontinuities of the indurated units were widely spaced, showed medium persistence and wide aperture, and were slightly weathered. Infill comprised predominantly loosely packed, very strong, coarse gravel. UCS of scoriaceous andesite and altered andesite was 25 5 MPa and 15 1 MPa respectively, allowing categorisation as very weak rock. Discontinuities of scoriaceous andesite were widely spaced, showed high persistence and wide aperture, and were moderately weathered. Discontinuities of the altered andesite were moderately spaced, showed low persistence and wide aperture, and were highly weathered. Infill of scoriaceous and altered andesite was loosely packed, moist, weak to very weak medium gravel. The block and ash flow was a poorly sorted, loosely packed, sandy, gravely and cobble rich matrix supported deposit. The Matahina Ignimbrite was a very weak, discontinuity-poor deposit. Shear box testing indicated cohesion and friction angle of 0 MPa and 42.1 (block and ash flow) and 1.4 x 10-3 MPa and 41.7 (Matahina Ignimbrite) respectively. These values are similar to published values. Correlation of each lithotechnical unit to its respective rock mass description site allowed approximate boundaries of each unit to be mapped. Each unit's mass strength was combined with measured bulk densities and incorporated into two dimensional slope profiles using the stability modelling package GalenaTM. Ten slope profiles of Putauaki were constructed. Failure surfaces for each slope profile were defined using the Bishop simplified multiple analysis method. Four slope profiles showed the potential for small scale failure (less than 0.1 km2 of material). The remaining six slope profiles showed the potential for large scale failure (greater than 0.1 km2 of material). Stability of these six slope profiles was investigated further in relation to earthquake force, watertable elevation, and a disturbance factor of the rock mass (D). Conditions of failure graphs for profile 6a showed that at low D (less than 0.4), earthquake forces and watertable elevation must be unrealistically high for the region (greater than 0.33 g; greater than 15% watertable elevation) in order produce a factor of safety less than 1. The remaining five slope profiles showed potential to be unstable under realistic earthquake forces and watertable elevations. Two of these profiles were unable to achieve stability at D greater than 0.8 (profile 4) and D greater than 0.9 (profile 5). A D value of 0.6 (intermediate between 0.4 and 0.8) is argued to most realistically represent Putauaki. The fact that Putauaki has not undergone large scale failure to date supports the conclusion that the constructed models overestimate the influence of those factors which promote slope instability. Maximum and minimum landslide run-out zones were constructed for the slope profiles exhibiting the potential for large scale failure. Definition of the position and extent of maximum and minimum run-out zones assumed H/L (fall height to run-out length) ratios of 0.09 and 0.18 respectively, as well as the 'credible flow path' concept. Identified impacts of landslides sourced from Putauaki include inundation of Kawerau Township, Tarawera River, forestry operations, road networks, and power supplies. Based on these impacts, the risk posed by landslides from each slope profile was categorised as ranging from relatively low to relatively high. Landslides sourced from the south-west flanks pose a relatively low risk due to their prerequisite of unrealistically high watertable elevations and earthquake forces. Landslides sourced from the north-west flanks pose a relatively high risk as minimum run-out will inundate north-east parts of Kawerau Township. Landslides sourced from the eastern flanks pose a moderate risk due to their run-out zones avoiding Kawerau Township.

Les glissements de terrain se produisent dans toutes les chaînes de montagnes où la résistance de massifs rocheux est insuffisante pour contrer l’action de la gravité. Les terrains volcaniques sont particulièrement susceptibles de s’effondrer car les édifices sont composés des lithologies diverses et variées qui peuvent être fortement fracturées. En plus, la croissance rapide des édifices volcaniques favorise leur instabilité et leur effondrement. L’activité magmatique est un facteur additionnel responsable de la déformation des édifices, tandis que l’activité hydrothermale réduit la résistance des roches volcaniques. Pour ces raisons, l’évaluation des aléas liés à l’effondrement des édifices et à la formation des avalanches des débris volcaniques mérite une attention particulière. Les caractéristiques physiques des composants des avalanches des débris ont une influence directe sur la dynamique de ce type d'écoulement. Les dépôts des avalanches de débris présentent une morphologie de surface composée des nombreuses collines (hummocks), qui montrent fréquemment les séquences volcaniques initiales, ce qui suggère un mécanisme de mise en place proche de celui des glissements de terrain. Cependant, d’autres dépôts présentent des crêtes allongées (rides) dont le mécanisme de formation est encore méconnu. Le volcan Tutucapa (sud du Pérou) a été affecté récemment par deux avalanches de débris. La plus ancienne, « Azufre », est d’âge Holocène et résulte de l’effondrement d’un complexe des dômes et d’une séquence volcanique altérée (hydrothermalisée) sous-jacente. La deuxième avalanche, « Paipatja », a eu lieu il y a seulement 200-230 ans BP et est associée à une grande éruption explosive du Tutupaca. Les dépôts de cette avalanche présentent notamment de nombreuses rides. Les deux dépôts d’avalanche montrent deux unités différentes : une unité inférieure, caractérisée par la présence des blocs altérés (hydrothermalisés) provenant de l’édifice basal, tandis que l’unité supérieure est constituée par des blocs du complexe de dômes actifs. Le travail de terrain montre que les rides de l’avalanche « Paipatja » présentent une forte variation de granulométrie entre leur partie centrale (enrichie en blocs grossiers) et leurs parties latérales, ce qui suggère un processus de ségrégation granulaire. Des expériences analogiques montrent que des écoulements de mélanges de particules des différentes tailles subissent un processus de ségrégation et de digitation granulaire qui engendre des rides par jonction de levées statiques qui délimitent un chenal d’écoulement. Le processus de formation des rides est facilité par de faibles différence de taille des particules dans des mélanges bidisperses. Ces résultats suggèrent que les rides observées au Tutupaca résultent d’un écoulement granulaire. Les principales caractéristiques morphologiques des structures formées lors de ces expériences de laboratoire ont été comparées qualitativement avec les structures observées dans les dépôts du Tutupaca. Les structures observées au Tutupaca montrent que deux mécanismes de mise en place peuvent coexister dans les avalanches de débris volcaniques : le glissement de blocs plus ou moins cohérents, et l’écoulement semblable à celui d’un matériau granulaire. Cela dépend probablement de la nature des différents matériaux à la source des avalanches. Cette information doit être prise en compte pour l’évaluation des aléas liés aux avalanches des débris car des mécanismes d’écoulement différents peuvent induire des fortes variations de la distance parcourue par ces avalanches. / Landslides occur in all mountainous terrain, where the rock strength is unable to support topographic loading. Volcanic rocks are particularly landslide prone, as they mix strong and weak lithologies and are highly pre-fractured. Also, volcanoes themselves, are peculiar mountains, as they grow, thus creating their own topographic instability. Magmatic activity also deforms the edifice, and hydrothermal activity reduces strength. For all these reasons, volcanoes need close consideration for hazards, especially for the landslide-derived rock avalanches. The characteristics and properties of different debris avalanche components influence their behavior during motion. Deposits are generally hummocky, preserving original layering, which indicates a slide-type emplacement. However, some deposits have ridged morphology for which the formation mechanisms are not well understood. Two recent debris avalanches occurred at the Tutupaca volcano (S Peru). The first one, “Azufre” is Holocene and involved the collapse of active domes and underlying older hydrothermally altered rocks. The second debris avalanche, “Paipatja” occurred 200-230 y BP and is associated with a large explosive event and this deposit is ridged. The excellent conservation state of the deposits and surface structures allows a comprehensive analysis of the ridges. Both deposits have two contrasting units: a lower basal edifice-derived hydrothermally-rich subunit and an upper dome-derived block-rich unit. Detailed fieldwork has shown that Paipatja ridges have coarser core material and are finer in troughs, suggesting grain size segregation. Using analog experiments, the process that allow ridge formation are explored. We find that the mixtures undergo granular segregation and differential flow that create fingering that forms ridges by junction of static léeves defining a channel flow. Granular segregation and fingering are favored by small particle size contrast during bi-dispersed flow. The results suggest that the ridges observed at Tutupaca are product of a granular flow We extract the morphological characteristics of the deposits of granular flows generated in the laboratory and make a qualitative comparison with the Tutupaca deposits. The description of the different landslide and debris avalanche features at Tutupaca shows that two types of debris avalanche motion can occur in volcanic debris avalanches: the sliding of blocks more or less coherent and a flow similar to a granular material. This probably depends on source materials and the conditions of different parts of the initial landslide. Such information should be taken into account when estimating hazards at other volcanic landslide sites, as the different behaviors may result in different run outs.