The mechanics of collapse calderas and landslides in stratovolcanoes and basaltic edifices John Browning* Supervised by Agust Gudmundsson and Phillip Meredith (UCL) Department of Earth Sciences, Royal Holloway, University of London, Egham TW20 0EX *(john.browning.2012@live.rhul.ac.uk) 3. Project aims 4. Future directions References: 1. Kilburn, C.J and Maguire, B.V.W (2001). Italian volcanoes. Terra Publishing, Harpenden 2. Macdonald, G.A (1972). Volcanoes. Prentice-Hall, New Jersey 3. Gudmundsson, A (2008). Toughness and failure of volcanic edifices. Tectonophysics. 471. 27-35 4. Hulme, G (1974). The interpretation of lava flow morphology. Journal of Geophysical research. 39, 361-383 5. Geyer, A and Marti, J (2008). The new worldwide collapse caldera database (CCDB). JVGR. 175 334-354 6. Lipman, P.W (1997). Subsidence of ash-flow calderas: relation to caldera size andchamber geometry. Bull Volc. 59 7. Gundmundsson, A. (2008). Magma-chamber geometry, fluid transport, local stresses and rock behaviour during collapse caldera formation 8. Gudmundsson, A and Nilsen, K (2006). Ring-faults in composite volcanoes. Geological Society, Special Pub. 269. 83-108 9. Gudmundsson, A (2011) Deflection of dykes into sills at discontinuities.Tectonophysics 500. 50-64 1. Introduction Volcanic structures can be simply grouped into two main types (fig. 1) : • Composite edifices or stratovolcanoes are tall, strong structures with slopes as steep as 35-42 o [1] . A stratovolcano is composed of different types of strata: lava flows, pyroclastic and sedimentary units and intrusions. • Basaltic edifices in contrast are usually gently slopping, for example the Hawaiian shield volcanoes slope between 2 and 12 o [2] . Basaltic lava flows are the most common eruptive material, with pyroclastics contributing only a minor part (~1%) 2 in the Hawaiian case. Figure 1. (left) the Hawaiian shield volcano of Mauna Loa , a basaltic edifice comprised of predominantly lava flows. As such the edifice can be seen as relatively homogeneous in terms of material properties . (Right) Teidi, Tenerife . A steep stratovolcano which has formed inside the Las Canadas caldera complex. The rock layers, units and contacts that constitute a stratovolcano have widely different mechanical properties. • How a stratovolcano can maintain such steep slopes is not well defined. • The most common explanation relates to the difference in viscosity and yield strength of the erupted products. However, based on the results of [4] it is clear that this is not an appropriate solution. • Instead it has been proposed [3] that a stratovolcano maintains steep slopes due to its considerable resistance to fracture propagation 2. Current understanding of caldera formation Figure 2. Four recently formed collapse calderas (top left) Piton de la Fournaise, 2007 (top right) Miyakejima, 2000 (bottom left) Mt Pinatubo, 1991 (bottom right) Fernandina – 1968. All examples of basaltic edifices experiencing small eruptive activity, the exception here being Mt Pinatubo. Source (CCDB) The aim of this project is to understand how much of the difference in collapse frequencies between composite and basaltic edifices can be explained in terms of difference in material toughness (fig. 6). A difference that is partly due to variation in lithologies and contact properties (fig. 5) and, therefore, related to the compositional range of magmas in the edifices. The problem of caldera formation has received much attention in recent decades, see [5]. Such work has allowed the creation of the Collapse Caldera Database [6] (fig. 2). However, there are several aspects regarding the frequency of collapse events with respect to eruptions that are still not well understood: • Unrest periods and eruptions within calderas are relatively common, however slip on existing ring faults is rare. • Many large caldera collapses have occurred with small or no eruptive volumes • Collapse events appear to be more frequent in basaltic edifices. It is argued that the mechanics of caldera collapse can only be understood through the local stress field of the host volcano at the time of ring-fault formation [8]. Several models are used to explain the process of caldera formation, perhaps the most favoured in its simplicity is the underpressure model (fig. 4) . Figure 5. A simple numerical model showing an overpressured (10 MPa) dyke approaching a tough (Youngs modulus = 100 GPa) layer [9]. Here the stress field rotates, and no longer supports propagation. In this case the dyke would be come arrested or turn into a sill, as can be witnessed in many examples of sills and dykes in Iceland (right). Figure 6. Fracture propagation is also influenced by material toughness, here a fault becomes arrested upon meeting a weak layer in a composite edifice, whereas stress field homogenisation promotes propagation in the basaltic edifice Figure 4. The underpressure or withdrawal of magmatic support model. Here an eruption causes a magma chamber to be partially drained and therefore underpressured, i.e less than lithostatic. A steeply dipping normally orientated bounding ring-fault forms allowing the collapse of the edifice into the chamber void. Knowing the stress conditions within a volcanic edifice is an important tool for predicting the likelihood of a volcanic eruption. It is hoped that the results from this study will be useful for assessing the likelihood of the formation of collapse calderas and large landslides during periods of unrest in volcanic edifices. Figure 3. A plate or piston caldera type, bounded by steeply dipping ring faults. Figure 7. Licancabur volcano in Chile. Measurements of the diameters and maximum heights of volcanic edifices in different tectonic settings using Google earth will be taken. A database of volcanic edifices will be created which will be used to better understand the growth and evolution of volcanic structures from different magma compositions. An initial study of South American volcanoes has begun to asses the feasibility of the method. Figure 8. Results from two simple 2D FEM numerical models using COMSOL. (left) A sill like magma chamber 8 km in diameter and 2 km in height, at around 4km depth with an internal pressure less than lithostatic, i.e underpressure in this case of 5 MPa. (right) Doming of a volcanic field, 10 MPa in this case. The graphs below show shear and tensile stresses at the near surface. A condition for ring fault initiation is that maximum tensile stress ( 3 ) should occur at the surface above the lateral ends of an associated magma chamber. This condition is satisfied in the model with doming, underpressure alone does not produce stress conditions favourable for ring fault initiation. [7] proposes a number of problems with the widely used underpressure model, perhaps the most salient are: • The poor correlation between collapse (caldera) volume and combined volumes of extrusive and intrusive material leaving the chamber. • The model requires a a fluid fracture (dyke) to remain open with a negative magmatic pressure. NEMOH network school 17-23 February 2013 0.5 km 1 km 1 km 1 km • A database of volcanic edifices from different tectonic settings will be created and used to understand structural growth and evolution (fig. 7) • FEM Numerical models will be created to assess the stress fields associated with edifice collapse, with particular emphasis on layers and contact properties (fig. 8) • RoĐk fraĐture edžperiŵeŶts will ďe ĐoŶduĐted at UCL’s RoĐk aŶd IĐe PhLJsiĐs Laď 0 1 2 3 4 5 6 7 8 9 10 0 5 10 15 20 25 30 35 Stress (MPa) Distance (km) Surface stresses Underpressure Shear Underpressure Tension 0 2 4 6 8 10 12 14 16 0 5 10 15 20 25 30 35 Distance (km) Doming Shear Doming Tension 10 MPa doming E = 40 GPa Homogeneous crust