On April 2013, a local scale seismic network, named OTRIONS, composed of twelve short period (1 Hz) three component seismometers, has been located in the northern part of the Apulia (southern Italy). In the first two months of data acquisition, the network recorded about one hundred very small (ML<2) magnitude earthquakes. A three-layer 1D VP velocity model was preliminarily computed, using the recordings of earthquakes occurred in the area in the period 2006-2012 and recorded by the national seismic network of INGV (Istituto Nazionale di Geofisica e Vulcanologia). This model was calibrated by means of a multi-scale approach, based on a global search of the minimum misfit between observed and theoretical travel times. At each step of the inversion, a grid-search technique was implemented to infer the elastic properties of the layers, by using HYPO71 to compute the forward models. In a further step, we used P and S travel times of both INGV and OTRIONS events to infer a minimum 1D VP velocity model, using a classical linearized inversion approach. Owing to the relatively small number of data and poor coverage of the area, in the inversion procedure, the VP/VS ratio was fixed to 1.82, as inferred from a modified Wadati diagram. The final 1D velocity model was obtained by averaging the inversion results arising from nine different initial velocity models. The inferred VP velocity model shows a gradual increase of P wave velocity with increasing the depth. The model is well constrained by data until to a depth of about 25-30 km.

We propose a model to explain lava tube formation by the growth of crust slabs from the levees of a channel toward its center, involving solid surface fragments. The flow dynamics are described with the steady state solution of the Navier-Stokes equation in a rectangular channel, for a Newtonian, homogeneous, isotropic and incompressible fluid. The cooling of a lava flow is described by the mechanism of heat conduction into the atmosphere. The presence of levees is taken into account both in the dynamical and in the thermal model. As long as the flow cools, a solid layer forms at the surface, much thicker near the levees than at the center of the channel. The crust that has formed breaks under the effects of the applied stresses. Solid blocks slow the fluid down, the shear stress at the interface between crust and fluid lava increases and the flow thickens. Using the thin elastic plate approximation, we determine conditions allowing a crust to resist both the action of the shear stress due to the drag of the underlying fluid and the tensile stress due to the weight of the crust itself, detecting the required crust thickness and distances from the eruptive vent where the tube can form.

Recent laboratory studies on the rheology of lava samples from different volcanic areas have highlighted that the apparent viscosity depends on a power of the strain rate. Several authors agree in attributing this dependence to the crystal content of the sample and to temperature. Starting from these results, in this paper we studied the effect of a power law rheology on a gravity-driven lava flow. The equation of motion is nonlinear in the diffusion term, and an analytical solution does not seem to be possible. The finite volume method has been applied to solve numerically the equation governing the fully developed laminar flow of a power law non-Newtonian fluid in an inclined rectangular channel. The convergence, the stability, and the order of approximation were tested for the Newtonian rheology case, comparing the numerical solution with the available analytical solution. Results indicate that the assumption on the rheology, whether linear or nonlinear, strongly affects the velocity and/or the thickness of the lava channel both for channels with fixed geometry and for channels with constant flow rate. Results on channels with fixed geometry are confirmed by some simulations for real lava channels. Finally, the study of the Reynolds number indicates that gravity-driven lava channel flows are always in laminar regime, except for strongly nonlinear pseudoplastic fluids with low fluid consistency and at high slopes.

We studied the conditions of crust and tube formation of a lava flow moving under the effect of gravity in a rectangular cross-section channel and assumed a power-law rheology for lava. We followed the work of Valerio et al. (2008), who studied the effect of surface cooling on the formation and accretion of the crust in the central region of the channel, assuming for lava a Newtonian rheology. According to these authors, tube formation is influenced by topography and channel morphology. In this work, we extended this study to a non-Newtonian rheology, in particular to the power-law rheology. Results indicate that a power-law rheology strongly influences the condition of crust formation but does not produce significant differences as a function of topographical or morphological variations.

We present here the results from dynamical and thermal models that describe a channeled lava flow as it cools by radiation. In particular, the effects of power-law rheology and of the presence of bends in the flow are considered, as well as the formation of surface crust and lava tubes. On the basis of the thermal models, we analyze the assumptions implicit in the currently used formulae for evaluation of lava flow rates from satellite thermal imagery. Assuming a steady flow down an inclined rectangular channel, we solve numerically the equation of motion by the finite-volume method and a classical iterative solution. Our results show that the use of power-law rheology results in relevant differences in the average velocity and volume flow rate with respect to Newtonian rheology. Crust formation is strongly influenced by power-law rheology; in particular, the growth rate and the velocity profile inside the channel are strongly modified. In addition, channel curvature affects the flow dynamics and surface morphology. The size and shape of surface solid plates are controlled by competition between the shear stress and the crust yield strength: the degree of crust cover of the channel is studied as a function of the curvature. Simple formulae are currently used to relate the lava flow rate to the energy radiated by the lava flow as inferred from satellite thermal imagery. Such formulae are based on a specific model, and consequently, their validity is subject to the model assumptions. An analysis of these assumptions reveals that the current use of such formulae is not consistent with the model.

Bends in lava channels are often observed in volcanic fields. The curvature of a channel affects flow dynamics and surface morphology and may be a trigger for the formation of lava tube. We propose a model to describe the effects of curvature on velocity, shear stress and the formation of crust at the flow surface. Lava is described as a Newtonian, homogeneous, isotropic and incompressible fluid. The steady-state solution of the Navier-Stokes equation is found for a unidirectional flow, in cylindrical coordinates. The flow levees are described as arcs of concentric circumferences, with their centres in the origin of the coordinate system. Under the assumption that the gravity force has no radial component, in the bend the fluid moves parallel to the levees. The velocity is assumed to depend on the radial coordinate only. As an effect of curvature, velocity and shear stress are asymmetric with respect to the centre of the channel. The maximum of surface velocity is shifted toward the internal levee, and the shear stress has larger values close to the internal levee. This effect is greater for wider channels. Heat radiation and convection into the atmosphere are considered as the main cooling mechanisms and the temperature distribution along the channel is calculated. Crust formation at the flow surface is considered under the assumption that solid lava is a plastic body. The amount of crust coverage is mainly controlled by the channel width: narrow channels have a greater coverage than wide channels for a given radius of curvature. The effect of a bend is to favour the crust growth toward the internal levee, while the crust coverage toward the external levee decreases. The presence of a bend in a lava channel may favour the formation of a lava tube. The analytical solution will serve as a benchmark for numerical models. Understanding the mechanism of formation of lava tubes is crucial to the simulation of actual lava flows and to evaluation of the associated hazard.

An effusive volcanic eruption results from a sequence of different processes, such as the pressurization of a magma chamber, the propagation of a dyke and the flow of lava at the Earth' surface. The aim of this paper is to establish relationships between the different quantities describing such processes. We consider a spherical magma chamber filled with a low-viscosity magma and included in a homogeneous and isotropic elastic half-space. We assume that, as a result of the inflow of fresh magma or a phase transition, the pressure in the chamber increases slowly during a finite time interval. Assuming that the pressure increase is linear in time, we calculate the stress field generated in the surrounding medium considering the chamber as a centre of dilation. We assume that a vertical tensile fracture originates at the top of the magma chamber after the rock strength is exceeded. The fracture is assumed to propagate quasi-statically along a vertical plane, driven by the stress distribution: both the cases of positive and negative buoyancy force are considered. The problem is solved in two dimensions by considering the fracture as a tensile Somigliana dislocation and expanding the associated stress release into Chebyshev polynomials. The fracture may reach the Earth's surface or not, depending on the depth and radius of the magma chamber, the rate and duration of pressure increase, the rock and magma densities and the rock strength. When the fracture reaches the Earth's surface, we assume that it becomes a vertical conduit. Magma pours out from the vent, driven by the pressure gradient in the conduit. Under the assumption of laminar flow of a Newtonian fluid, we evaluate the initial effusion rate as a function of the relevant model parameters. The flow rate is found to be a non-linear function of the density contrast. We also establish a relationship between the flow rate in the conduit and the initial thickness of the ensuing lava flow, in the case of effusion on a steep slope.

The cooling of a lava flow, both in the transient and the steady state, is investigated considering that lava rheology is pseudoplastic and dependent on temperature. Lava exits from the vent with constant velocity and flows down a slope under the effect of gravity force inside a channel of rectangular cross section. We consider that cooling of lava is caused by thermal radiation into the atmosphere and thermal conduction at the channel walls and at the ground. The heat equation is solved numerically in a 3D computational domain and the solution is tested to evaluate the numerical errors. We study the steady state and the initial transient period of lava cooling. Results indicate that the advective heat transport significantly modifies the cooling rate of lava slowing down the cooling process. Since the lava velocity depends on temperature, the cooling rate depends on the effusion temperature. Velocity profiles are modified during cooling showing two marginal static zones where the crust can form and remain stable. The fraction of crust coverage is calculated under the assumption that the solid lava is a plastic body with temperature dependent yield strength. We numerically confirm that heat advection can not be neglected in the mechanism of formation of lava tubes.

In this work we studied the effect of a power-law rheology on a gravity driven lava flow. Assuming a viscous fluid with constant temperature and constant density and assuming a steady flow in an inclined rectangular channel, the equation of the motion is solved by the finite volume method and a classical iterative solutor. Comparisons with observed channeled lava flows indicate that the assumption of the power-law rheology causes relevant differences in average velocity and volume flow rate with respect to the Newtonian rheology.

Nell’ambito del progetto OTRIONS è stata installata, sul promontorio del Gargano, una rete sismica di 12 sismografi a corto periodo (Lennartz LE-1DV,1Hz) a tre componenti, a partire dal 1° Aprile 2013. L’area ricoperta dalla rete sismica ha una estensione di circa 30 x 30 km2 e racchiude i comuni di S. Giovanni Rotondo, Lucera e Manfredonia. Presso l’Università di Bari è stato realizzato un laboratorio sismico, dove i dati vengono ricevuti e processati in tempo reale, mediante l’uso del software SeisComp3 prodotto da GFZ-Potsdam. Nei primi due mesi di attività della rete, sono stati registrati circa un centinaio di terremoti, caratterizzati da una magnitudo inferiore o uguale a 2. Adottando un metodo di tipo Montecarlo (Hypo 71 e Velest), è stato ricavato un modello di velocità 1D a 3 strati delle onde sismiche, dall’inversione dei tempi di arrivo registrati nel Gargano dall’INGV nel periodo 2006-2012. Tale modello è stato ottenuto, mediante un approccio multi-scala, basato sulla ricerca globale del misfit minimo tra tempi di arrivo osservati e sintetici. Per ogni inversione, è stata adottata una tecnica di grid-search per determinare le proprietà elastiche degli strati del sottosuolo: i nuovi valori di velocità sono stati usati come parametri iniziali da dare in input ad Hypo 71, per ottenere in output modelli ulteriormente dettagliati. Successivamente sono stati utilizzati i tempi di arrivo delle fasi P ed S, di eventi registrati dall’INGV e da OTRIONS, per definire un modello di velocità 1D, mediante un approccio di inversione classico linearizzato (Velest, Kissling et al. 1994). E’ stato necessario fissare il valore del rapporto Vp/Vs a 1.82, come ottenuto dal diagramma di Wadati modificato, a causa del numero relativamente piccolo di dati e di una bassa copertura azimutale. Per la stessa ragione i ritardi di stazione non sono stati determinati dall’inversione ma fissati sulla base delle conoscenze geologiche dell’area. Il modello finale 1D di velocità è stato ricavato dalla media dei risultati ottenuti, usando nove diversi modelli iniziali di velocità. Il modello dedotto per la velocità delle onde P, mostra un graduale aumento di velocità al crescere della profondità e risulta ben vincolato ai dati fino ad una profondità di circa 25-30 km. In questa nota verranno presentati i risultati di questo studio e le localizzazioni di oltre 1100 eventi registrati dalla rete mista OTRIONS-INGV nel periodo aprile 2013- marzo 2014.

A partire dal 1° aprile 2013, nell’ambito del progetto OTRIONS è stata installata, sul promontorio del Gargano, una rete sismica di 12 sismografi a corto periodo (Lennartz LE-1DV,1Hz) a tre componenti. L’area ricoperta dalla rete sismica ha una estensione di circa 30 x 30 km e racchiude i comuni di S. Giovanni Rotondo, Lucera e Manfredonia. Le stazioni sismiche inviano i dati in tempo reale ad un laboratorio situato all’Università di Bari, attraverso un protocollo seed-link e vengono analizzati mediante il software SeisComp III prodotto da GFZ-Potsdam. Tale software consente di localizzare in tempo reale gli eventi sismici. Nel periodo 24 Aprile 2013 - 23 Giugno 2013 la rete ha registrato 67 eventi, la maggior parte dei quali sono stati localizzati nel Gargano. Dall’analisi della sismicità registrata nel Gargano dall’INGV nel periodo 2006-2012 e dai dati acquisiti dalla rete OTRIONS è stato ricavato un primo modello di velocità 1D delle onde sismiche nel sottosuolo dell’area, usando un approccio di tipo Montecarlo basato sull’uso in cascata di Hypo 71 e Velest. Inoltre, utilizzando le forme d’onda meno rumorose, è stata eseguita la determinazione del momento sismico, della frequenza d’angolo, della dimensione delle rotture e del fattore di qualità delle onde P dall’inversione degli spettri delle onde P. In questa nota verranno presentati i risultati preliminari di questo studio.

We investigated the cooling of a lava flow in the steady state considering that lava rheology is pseudoplastic and dependent on temperature. We consider that cooling of the lava is caused by thermal radiation at the surface into the atmosphere and thermal conduction at the channel walls and at the ground. The heat equation is solved numerically in a 3D computational domain. The fraction of crust coverage is calculated under the assumption that the solid lava is a plastic body with temperature dependent yield strength. We applied the results to the Mauna Loa (1984) lava flow. Results indicate that the advective heat transport significantly modifies the cooling rate of lava slowing down the cooling process also for gentle slope. Progress in Industrial Mathematics at ECMI 2012Progress in Industrial Mathematics at ECMI 2012 Look Inside Other actions Reprints and Permissions Export citation About this Book Add to Papers Share Share this content on Facebook Share this content on Twitter Share this content on LinkedIn